Академический Документы
Профессиональный Документы
Культура Документы
Subnuclear Physics
THE SUBNUCLEAR SERIES
7. 1969 SUBNUCLEARPHENOMENA
Volume 1 was published by W. A. Benjamin, Inc., New York; 2-8 and 11-12 by Academic Press,
New York and London; 9-10 by Editrice Compositori, Bologna; 1315 by Plenum Press, New York
and London.
The lIflys of
Subnucleor Physics
Edited by
Antonino Zichichi
European Physical Society
Geneva, Switzerland
A. zichichi
Geneva, September 1978
v
Pn06eh~on Sidn~y Coleman h~ b~~n awand~d ~h~
P~z~ 06
"Beh~ L~c~lUt~n"
vii
viii CONTENTS
Index................................................... 1231
THE WHYs OF SUBNUCLEAR PHYSICS
Antonino Zichichi
CERN
Geneva, Switzerland
of the new lepton with 'V l.9 GeV mass, of the "new physics" in the
(3.9-4.5) GeV mass range of SPEAR, and of the dileptons plus the
electron-strange particle production induced by high-energy v .
]J
1. DO WE UNDERSTAND ISOSPIN?
2. DO WE UNDERSTAND STRANGENESS?
:::: 2-3 mb 0
TIp
~ 3.1 GeV m
TI
ii) Why are the weak non-leptonic decay rates three orders of
magnitude greater than the weak leptonic decay, i.e.
A -+ S
?
A -+ all
vi) The charm yield at 4.415 GeV is 'V 1/3 that at 4.0.28 GeV.
Is this due to charm burning?
i) \lliy is the Cabibbo angle 'V 2o.O? Can the answer be found
v~a spontaneous symmetry breaking?
iii) \lliy do _
m(wa)
_ and .:::..:.:.:..L
m(w ) not coincide with the observed mas-
ses vector mesons? Is this because of SU(3) breaking?
4 A. ZICHICHI
5. DO WE UNDERSTAND LEPTONS?
ii) Another heavy lepton of the standard type with its own
neutrino -- seems to show up, with standard electromag-
netic properties, at SPEAR. Why another lepton, and so
heavy,
m ~ 1.9 GeV/c 2 ?
(~~) C)
( (:)
iii) lfuy is
o(PP)elastic I
o(PP)total
" 5
when w -+ 1 ?
iii) Why is
a(7Tp) 2
?
a(pp) "3
Is this because mesons are made of two quarks and baryons
of three quarks?
***
Let us go into deeper problems.
i) Why does
K -+ 2n
L
go? Or more generally, why ~s CP violated? Is this
because a new interaction -- superweak -- is at work?
iii) Why are the proton, the neutron, the AD, as examples of
baryonic states, and the e-, ~-, v e ' as examples of lep-
tonic states, all left-handed when they interact weakly?
Is this due to the fact that they sometimes transform
from one to the other? Is this related to the fact that
the bare electric charges of the electron and proton
(i.e. of the two best measured so far) are equal?
~ -m~ = ~* - m~
which is checked by experiment, as well as the Johnson-
Treiman relations for (TIp) and (Kp) total cross-sections,
and the relations between decay amplitudes in non-leptonic
strange particle decays.
i) Is all this because unitary spin and Dirac spins are cor-
related?
iii) The proton lifetiITle is"T 'V lei 0"T. at leas t. Why
P unlverse
is the proton so stable?
* * *
In spite of the classification into many groups~ many problems
are obviously correlated. This is again a crucial point of inves-
tigation which will be developed in the years to come. The above
series of problems~ some very hot and fashionable~ others well-
known since so long that they are even forgotten~ testify the vital
role that Subnuclear Physics has in present day scientific know-
ledge.
WHY IS THERE CHARM, STRANGENESS, COLOUR AND ALL THAT?
Harry J. Lipkin*)
1. INTRODUCTION
The neutron and the proton have similar masses and strong in-
teractions. They have different electric charges and electromag-
netic interactions. The similarity of their strong interactions is
expressed formally by the principle of charge independence of nuclear
forces and by the symmetry of isospin invariance. The symmetry is
broken by the electromagnetic interactions which do not conserve iso-
spin but only the z component or electric charge. This symmetry
11
12 H. LIPKIN
pendently of the validity of the quark model. This raises the ques-
tion: Why do hadrons have Abelian and non-Abelian quantum numbers
which suggest that they are made of quarks when quarks are not ob-
served as free particles in nature?
Table 1.1
Additive quantum numbers and non-Abelian symmetries
Table 1.2
Symmetry algebras in physics
Laboratory
Space time Hilbert space
multiplet structure
conservation laws operator algebras
of spectrum
Rotations
Px ... Px cos e+ Py sin e 1/10. ... L Co.[3 1/1[3
[3 _J,M J n, n + i
OJ M = -j, +J
Conservation of angular -
momentum [J , J J= iJ ; [J , Jz
x y z
Isospin transformations
also finds operators like the angular momentum operators which gene-
rate the symmetry transformations. The commutation relations among
these operators generate an algebra. Analysis of the algebra leads
to new operators like J2 which commute with all of the generators
and determine the structure of the mUltiplets whose states transform
into one another under the algebra.
The answer in this case is a model in which all complex nuclei are
made from an elementary doublet building block, the nucleon, if the
forces which bind nucleons together to make nuclei are charge inde-
pendent.
Soon after the SU(3) symmetry came SU(6) which followed from
the observation that SU(3) multiplets with different spins fit to-
gether into supermultiplets of the SU(6) algebra, as one would ex-
pect for composite models with basic building blocks having three
flavours and spin one half.
18 H. LIPKIN
by pushing all states which are not colour singlets up very high in
mass or throwing them out of Hilbert space all together. The latter
is the limit of pushing them up in mass to the point where they have
infinite mass. A search for a dynamical theory described in space
time which would have these properties in Hilbert space led to non-
Abelian gauge theories which depressed all colour singlet states and
might lead to quark confinement, and the pushing up of all non-
singlet states to infinite energy. This happens exactly in a I + I
.. 9) . . . .
d~mens~onal model ,where a quark-ant~quark pa~r are l~ke a pa~r of
condenser plates and separating the plates requires infinite energy.
13)
2. STRANGENESS, CHARM AND MASS SPLITTINGS
One starts in the laboratory by noting that pions and kaons have
different masses, and that additional strangeness goes with increas-
ing mass. By analogy with the breaking of rotational invariance
with a magnetic field that transforms like a vector under rotations,
one can assume that the breaking of SU(3) symmetry transforms like
the SU(3) analogue of a vector, namely an octet. This gives the
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 21
Q L l,Q
l<-Tt=I<~.. ~=r:~A. S ... r: (2.1)
where the L above the equality implies that linear masses should be
used and the Q above the equality implies that quadratic masses
should be used.
G L.
D -1T::: l>*-e = C:-~ (2.2)
where the last equality is left open s~nce the doubly charmed baryon
analogous to the ~ has not yet been found. This formula also agrees
with experiment, as shown in Table 2.1. Thus changing a non-strange
quark in the p to a strange or to a charmed quark produces a linear
mass shift which is equal to that produced by the corresponding change
of a quark in the ~, while the shift in squared mass is equal to that
produced by the corresponding quark change in the pion.
(2.3)
(2.4)
15)
Since the present experimental information on charmed baryons
gives a mass of 2260 for the Co and a mass of 2500 for a broad peak
interpreted to be the unresolved Cl - Cl* combination, it is conveni-
ent to rewrite Eq. (2.4) as
(2.5)
Table 2.1
a) Strangeness splittings
Q L L.Q
K - IT K* - P
~ (GeV) 0.35 0.12 0.15 0.12
t.M2 (GeV) 2 0.22 0.20
b) Charm splittings
Q L
D- IT D* - P C* - t.
17)
3. HIGH ENERGY SPECTROSCOPY
3.1 Introduction
? Nobody has thought It will be found; it's Who knows? The theorists have not
of it there thought of it yet
t.)
til
26 H. LIPKIN
Resonances with masses in the several GeV range have very many
open decay channels. Their branching ratios into anyone exclusive
channel are of the order of 0.1%. Since the signature for the detec-
tion of such a resonance generally picks a particular decay mode, the
signal is proportional to the branching ratio and is very small. The
crucial factor in discovering and confirming such high-mass resonances
is the signal-to-noise ratio.
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 29
where a(P+x) and a(T+X) denote the cross-sections inclusive for pro-
duction of the particle P and the trigger T in the reaction under
consideration and BR(T) denotes the branching ratio for the appear-
ance of the trigger T in the decay of the particle P.
For states which do not have the quantum numbers of the photon
or of the meson-baryon. nucleon-nucleon or nucleon-antinucleon sys-
tem. some possibilities exist for production via the decays of states
which do have these quantum numbers; e.g. in the production of the
positive parity charmonium states by radiative decay of the ~' and
the production of charmed particle paris by the decays of higher vec-
tor resonances.
For states not easily produced in this way and available only ~~
Single ~ spectroscopy:
K * -+ K + <P (3.2a)
II. * -+ II. + ~ (3.2b)
L: * -+ L: + ~ (3.2c)
-* -+ ::: + ~
-
(3.2d)
r2 * -+ r2 + ~ (3.2e)
~ * -+K + K + ~ (3.2f)
F -+'fT + ~ (3.2g)
F -+ leptons + ~ (3.2h)
-0
Fr -+ 'fT0 + ~ (3.2i)
AS + 'fT0 + ~ (3.2j)
'fT
cp~ spectroscopy
(3.3a)
3.2.3 ~~E~S!~~~_~iS~~!~!~~
spectrum, e.g., looking for a charmed baryon decaying into A3n. The
problem is how to use the angular distributions of these four par-
ticles in the centre-of-mass system of the four-particle cluster
(hopefully the rest system of the new particle) as a means of dis-
tinguishing between signal and background. Three axes are relevant
for examining the angular distributions: (1) the direction of the
incident beam momentum, (2) the direction of the momentum of the
four-particle clusters, and (3) the normal to the production plane.
Signatures which characterize the new particle appear most clearly
in angular distributions with respect to the direction of the mo-
mentum of the four-particle cluster or with respect to the production
plane. But signatures for the noise will show up in angular distri-
butions with respect to the incident beam direction.
4. QUARKONIUM SPECTROSCOPY 1 7)
(cC.iI=OJ.rP=~-)-+(9~j 1:0).rPI:~-}-+
(~cfj I::IfJ jP= 1f Pf )+ l (4.lb)
In case (a) the photon carries away its angular momentum and parity
before the OZI violation. and the violation occurs in a system having
the space-spin quantum numbers of the final state. In case (b) the
OZI violation occurs in a system having the space-spin quantum numbers
of the initial state before the photon carries away angular momentum
and parity. The photon can now carry away isospin zero or one. and
the final state can be both isoscalar and isovector. Thus the isospin
properties of the final state contain information on the space-spin
state in which the OZI violation occurred.
and only in the yy decay mode. There is interest in seeing the ha-
dronic decay modes, and any ingenious method for seeing such decay
modes with hadronic production would constitute a real breakthrough
in x-onium spectroscopy. If the estimate (4.2) of the hadronic pro-
duction cross-section is reasonable, there may be some hope for de-
tecting the ne via the decay mode after production in pp colli-
sions. The figure of merit for this process can be estimated by
+ -
comparison with the detection of the J in the e e decay mode.
showed no events, while the same run observed about 100 events of
+ -
J/~ + ~ ~. This is still consistent with the result (4.5) of equal
signal/noise and comparable signals for the two processes, because
the spectrometer had a much lower acceptance for 's than for muons.
The absence of any signal confirms that the background is low, and
that any further experiments with increased sensitivity might see a
small signal without appreciable background. Note that even three
events for at 2.8 GeV with no background would constitute serious
supporting evidence for the existence of the n , whereas several hun-
c
dred events in another decay mode against a background of thousands
of events would be ambiguous.
The nA2 decay mode of the nc has also been suggested as a pos-
sible useful signature 23 ). A detailed analysis of the hadronic de-
cays of the n has been given by Quigg and Rosner 37 )
c
38)
The recent beautiful experiment at DESY reported by Schopper
. .
show~ng ev~dence for the F and F* mesons 39) is an example of how
choosing an appropriate signature minimizes background and gives
serious evidence for these particles with only a few events. The
signature in this case was three photons and a pion, with one photon
having a low energy and the other two having the mass of the n.
44 H. LIPKIN
(4.6)
This would give three photons, one of 300 MeV (or less if the
nc is not at 2.8 GeV but higher) and the other two having the mass
of the n, and two additional charged pions.
S. COLOUR
Many reasons have been proposed for introducing colour, and not
all of them are compatible. Colour is needed by:
"0 )
1) People who like ordinary Fermi statistics for quarks and do
not like baryon models with three spin 1 quarks in symmetric rather
than in antisymmetric states.
2) .
People who l1ke .
1ntegral .charge
electr1c " 1)
bers must have the same properties. In other models quarks of dif-
ferent colours have different observable properties, e.g different
electric charges. This possibility has been used to construct models
with quarks of integral electric charges. 8uch integrally-charged
coloured quarks cannot satisfy the Gell-Mann-Nishijima relation and
must have non-zero eigenvalues of a new additive quantum number which
appears in the modified Gell-Mann-Nishijima formula. The electromag-
netic current then has a component which is an 8U(3)f singlet and
which is not a singlet in 8U(3) There is a definite conflict be-
c
tween the use of integral charges and the use of colour as an exact
symmetry of nature in a non-Abelian gauge theory. If quarks of
48 H. LIPKIN
Some insight into the coloured quark models is given by the ana-
logy of a world in which all low-lying nuclear states are made of
deuterons and have isospin zero, free nucleons have not yet been seen
and experiment has not yet attained energies higher than the deuteron
binding energy or the symmetry energy required to excite the first
I = 1 states. In this isoscalar world where all observed states have
isospin zero the isovector component of the electromagnetic current
would not be observed since it has vanishing matrix elements between
isoscalar states. The deuteron energy level spectrum (something like
that of a diatomic molecule) would indicate that the deuteron was a
two-body system, but there would be no way to distinguish between the
neutron and the proton. The deuteron would thus appear to be composed
of two identical objects which might be called nucleons. Since the
deuteron has electric charge +1, the nucleon would be assumed to have
electric charge +1. Furthermore, the nucleon would be observed to
have spin ~ and be expected to satisfy Fermi statistics. However,
the ground state of the deuteron and all other observed states would
be found to be symmetric in space and spin. Thus, the nucleon would
appear to be a spin 1 particle with fractional electric charge and
peculiar statistics.
Thus if each quark in the baryon attracts the antiquark, some addi-
tional mechanism must be found to prevent it from being bound to the
quark system.
Our three whys involve only the strong interactions which do not
depend upon the couplings of quarks to the electromagnetic and weak
currents. The following discussion thus applies to both fractionally
charged and integrally charged models.
(5.1a)
-+
where t. is the isospin of par,tiele i and V contains the dependence
~
(5.2)
(5.31'1)
The interaction (5.2) and the mass formula (S.3b) were first pro-
41) d . . .
pose d by Nambu , an the saturatLon propertLes of the LnteractLon
53)
were considered by Greenberg and Zwanziger However, the re-
markable properties of this interaction as demonstrated above in the
simplified example of the analogous deuteron world have received
little attention.
The formula (S.3b) can test the triality why or the meson-baryon
why by showing whether observable "zero mass" hadron states exist for
a given number of quarks and antiquarks. However, it cannot test the
exotics why, since it gives no information about the spatial proper-
ties of the states. It cannot distinguish between one-particle
states and multiparticle scattering states and all zero-triality ex-
otic states are allowed as multiparticle states.
(S.3c)
The model thus gives observable hadron states for all quark and anti-
quark configurations for which C =0 states exist. Since C =0
states exist only for configurations of triality zero, this answers
the triality why.
Table 5.1
Values of the interaction and mass parameters c-nc and c/c
(5.4a)
(5.4b)
where particles 1 and 2 are quarks, 3 and 4 are antiquarks and (ij)l
denotes that particles i and j are coupled to C = O. Several use-
ful identities follow from the properties of the C = 0 two-particle
state:
(9"cr+~3cr)lo(>=(~lcr+94G")J0(> = (5.5d)
(5.5e)
- -(8/a)<0(1~>= -8/9
By operating with the interaction (5.2) on the wave functions
(5.4) and eliminating the colour variables with the aid of the iden-
tities (5.5) we obtain
and
where
Solving the secular equation for Eqs. (5.6) gives the eigenvalues
for U.
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 61
attract and repel an external particle and the net force exactly
cancels. Thus theory and experiment now agree on the absence of
naive exotics. But the possibility exists of higher exotics. Mole-
cular-type exotics in which attraction results from spatial polari-
zation of one hadron by another have been considered, but the re-
sults (5.9) indicate that the force is insufficient to produce bind-
54)
ing. Rosner has postulated the existence of exotics from the
point of view of finite energy sum rules and duality. This approach
has been carried further by other theorists and experiments have
been suggested in a search for exotics by baryon exchange processes.
quantum numbers also exist, these were not taken seriously as pos-
sible configurations for the known states, because there was no good
theoretical reason why such states should be present and their exo-
tic partners should be absent. But now there seems to be evidence
that the low-lying 0++ nonet is indeed such a qqqq state~), and there
are new convincing theoretical reasons why only states with non-
exot1c quantum numb ers are seen 55)
o
(5.10)
(S.ll)
The colour singlet baryon is the scalar product of three quark vec-
64 H. LIPKIN
(5.13)
--
product of antisymmetric diquark vector and a quark vector
B=j)Q (5.14)
2M (5.15a)
There is also the state formed by coupling the two quarks and two
antiquarks each to an antisymmetric vector and coupling the two
vectors to a scalar
(S.15b)
-
BB .... X (S .16a)
One might imagine the situation where the baryonium state X created
in some reaction would prefer to decay into the baryon-antibaryon
state via the transition (S.16) rather than to decay into two mesons
by breaking up into two quark-antiquark pairs, because the latter
transition involves changing the colour couplings. In particular,
this situation could arise if there is an appreciable spatial sepa-
ration between the diquark and the antidiquark.
One can picture lines of force joining the quarks and anti-
quarks by analogy with electrodynamics but with essential modifi-
cations following from the non-Abelian character. A colour singlet
quark-antiquark pair would have lines of force originating on the
quark and ending on the antiquark as shown in Fig. S.la. Figure
S.lb shows the two-meson system described by Eq. (S.lSa) as two
such pairs with lines of force joining the members of each pair but
no lines of force connecting the two pairs. Figure S.2 shows the
baryon described by Eq. (S.12) as three quarks at the vertices of
a triangle with lines of force between them. Here the non-Abelian
nature has new effects, with lines joining each quark and its neigh-
bour rather than quark and antiquark, and with one line acting as
a source for another, since the lines themselves carry colour. The
coupling of the baryon described in Eq. (S.14) as the product of a
diquark and a quark is seen by cutting the baryon diagram to sepa-
rate a quark from a diquark and noting that the lines of force going
from a quark to an antiquark in a meson. This again shows us that
66 H. LIPKIN
~!}
Q
Baryon
~
Diquark
Fig. 5.2 Fig. 5.3
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 67
the diquark has the same quantum numbers in colour as the antiquark.
The diquark is thus an unsaturated system with lines of force j oin-
ing the two quarks but other lines of force left out and searching
for a partner as shown in Fig. 5.3. However, the number of lines
of force originating from such a diquark are not twice the number
originating from a quark but only the same as the number originating
from a quark. Here we see again the essential difference between
non-Abelian and Abelian vector theories. The lines of force for
the electron-positron system are very much like the lines of force
for quark-antiquark system. But the lines of force for the antisym-
metrized diquark system are very different from the lines of force
in the two electron system where there are no lines joining the two
electrons and the number of lines which are unsaturated and looking
for partners is exactly twice the number from one electron.
Vector
Q Vector Q
Baryon-Antibaryon Pair Baryonium
The two couplings (S.lSa) and (S.lSb) are not the most general
couplings to construct a scalar from four vectors. Simple analysis
shows that there are three independent couplings corresponding to
coupling any two vectors to a scalar, vector, or tensor, coupling
the other pair in the same way and coupling them both to a scalar.
However, when we return to the realistic case of complex vectors
and SU(3), there are only two independent couplings. The two
states (S.lSa) and (S.lSb) are 1inear1y.independent but not ortho-
gonal and constitute a complete non-orthogonal basis for colour
singlet state of the two-quark, two~antiquark configuration.
6.1 Introduction
like positronium, the forces saturate and the residual force be-
tween the two neutral systems is very small and does not produce a
state more strongly bound than the original two-particle states.
From the experimental observation that there is no strongly bound
doubly charged state of two positive pions, we conclude that the
pion is more like positronium than like the deuteron.
mass difference is much larger than the binding energy of the deu-
teron:
Jaffe has simply used the N-~ and p-n mass splittings as input
for the strength of the spin-dependent interaction and calculated
74 H. LIPKIN
For the qqqq system flavour antisynnnetry gives two very in-
55,59)
terest1ng qua11tat1ve pred1ct10ns
low the pn and the KK thresholds and its dominant decay mode nn is
ambiguous because of mixing in the n of both strange and non-strange
components. But striking features in decay rates should be seen in
the first four-quark isovector state which is above the pn and the
KK thresholds. An unusual decay pattern ~s seen for the tensor
meson TS with the quantum numbers of the A2 but which does not de-
n
cay into pn but rather into KK, KK
* and nn and for the axial vec-
tor meson AS with the quantum numbers of the B, but with the n de-
n
cay dominant and wn forbidden. The n decay mode is particularly
interesting, since it is forbidden for all normal quark-antiquark
mesons by the OZI rule, while perfectly allowed for four-quark
states. Thus a search for n resonances might be an interesting,
way to find four-quark mesons.
The observation that the 0 and the S* are scalar mesons lying
below the ~ which is the lowest vector meson in the (ss) configura-
tion has interesting implications for the new particle spectrum.
If we assume the charm-strange analogy and replace all strange
quarks in the 0, S* and ~
.the.
by charmed quarks, we pred1ct eX1st-
ence of isoscalar and isovector scalar charmonium states denoted by
oc and S* with configurations like (uu
- c~) which should lie below
c
the lowest (cc) vector meson state, namely the J/~. If we replace
*
only one strange quark in the 0, Sand by charmed quarks, we
predict the existence of exotic charmed-strange states with con-
figurations (udcs) which should lie below the lowest (cs) vector
*
meson state, namely the F. We now consider these possibilities in
more detail.
The four-quark states with four different flavours and the same
colour-spin couplings as the low-lying 0++ nonet constitute a set of
charmed-strange scalar mesons which are expected to lie in the same
mass range as the two-quark charmed-strange F mesons. These include
exotic states whose quantum numbers differ from those of the F by
having either the wrong sign of strangeness or the wrong isospin.
The two types of states are denoted by FI (udcs, etc. - wrong iso-
spin) and FS (udcs, etc. - wrong sign of strangeness). The "crypto-
exotic" (u~cs) four-quark state with the same quantum numbers as the
F is denoted by Fx'
-
The FI can be considered as an Fn or DK reso-
nance or bound state, the FS as a DK resonance or bound state, and
the F as an excited F coupled to the DK channel. One way to see
x
the relation of these exotics to the low-lying 0+ nonet is to note
that changing a charmed quark to a strange quark in the FI and Fx
gives a state in the 0+ nonet, while FS has no such charm-strange
analogue state. Rough estimates of their masses are near the DK
threshold. If the FS and Fx are below the DK threshold, as appears
likely, they would be stable against strong decays and decay only
weakly or electromagnetically.
80 H. LIPKIN
Table 6.1 lists these states with their quark structure, quan-
tum numbers, dominant strongly coupled channels, possible weak and
EM decay modes and their "charm-strange analogue" states in the
light quark spectrum, obtained by changing the charmed quark to a
strange quark.
~
82 H. LIPKIN
1 ""
M(F")+M(lt) ~ M(F"IqqcS) < M(K)+M(D) (6.2b)
(6.3a)
(6.3b)
(6.4a)
(6.4b)
(6.4c)
(6.6a)
(6.6b)
Note that the decay (6.6) can give only neutral kaons and not
charged kaons because the final states have strangeness 2 and zero
electric charge. Thus although these final states have lost the
memory of the double strangeness part of the memory remains in the
absence of the charged two-body kaon decay modes. Since nearly all
other non-leptonic decays into final states containing kaons tend
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 85
Because the same !inal state KSKS is produced both in the de-
cay os the FS and the FS ' there can be mixing of the two states as
in the neutral kaon system. In the approximation where CP viola-
tion is neglected, the mixing will lead to eigenstates of the mass
matrix which are CP eigenstates. The decay (6.6) will then be al-
lowed only for the eigenstate which is even under CP. The state
of odd CP will not be able to decay into two pseudoscalar mesons.
The other charge states of the FI would decay weakly. The multi-
body decays would resemble the expected decay of the F with an ex-
tra pion, and the Cabibbo favoured decays would be into states of
zero strangeness. In addition, there would be the two-body decays:
r::- ++
... - +
+-
rI .... 1t'- TC (6.8a)
(6.8b)
then the eigenstates will be CP eigenstates and the even state will
have the decay modes (6.8b) while the odd CP state will not decay
into two pseudoscalars and will decay to three or more, in the non-
leptonic modes and into semileptonic decay modes. Note that this
P~ - F~ mixing will be much stronger than KO - KO mixing in a gauge
theory, because it can go via exchange of two intermediate W bosons,
with all vertices Cabibbo favoured and no cancellation of the GIM
type.
one or more pions, without violating the OZI rule. One might ex-
pect a11 charmonium states which can decay hadronica11y to 0 plus
c
pions to be very wide. However, angular momentum and parity selec-
tion rules seriously restrict the states which are allowed to de-
cay into n and pions. The final state 0 n has unnatural parity
c c
and even G. The I = a states produced from a strong decay of
charmonium would then have positive C. Thus of all the low-lying
+
sand p wave charmonium states, only the nc ' n~ and the X(l ) would
be allowed to decay into 8 plus a pion.
c
The final state 8
c
= 2n has I = a only if the two pions are
in an odd I = 1, odd J state like the p. The final state 0
c
+ p
has odd G and could be produced in the decay of the ~'. It would
be interesting to see whether this decay exists in present data,
whether it is ruled out and if so, what the upper limits are on
this decay. The 8 could also be produced in a radiative decay of
c
the ~.
dc -+ Wr (6.9b)
~,"ew (6.9d)
'P'" l + de (6.lOb)
(6.lOc)
One might think that the electric charge of the quark is ob-
servable directly from measurements of the electromagnetic coup-
lings of hadrons as shown in Fig. 7.1 since these hadron couplings
are commonly assumed to be given
by the sums of the couplings of
the constituent quarks. However,
the coupling of the electromagne-
tic current to a colour singlet
hadron depends only on the colour
averaged quark charge <Q>; i.e.,
c
the average over colour of the
charges of quarks having identical
values for all other quantum num-
Fig. 7.1 Photon absorption by bers except colour. This can be
coloured quark parton seen because a colour singlet
a) by red quark state is completely symmetric in
b) by white quark
c) by blue quark colour space and can give no
90 H. LIPKIN
(7.la)
quark model, and the arguments 8 f ,lc denote that this current trans-
forms under flavour and colour like an octet and singlet, respec-
tively. In the Han-Nambu model, ~J is a flavour singlet and a co-
lour octet. Thus
(7.lb)
(7.lc)
94 H. LIPKIN
(7.ld)
+ yy. Decays of this type of meson into two photons are assumed to
. , 32)
be descr~bed by a tr~angle d~agram . We cons~der all poss~ble
(7.2a)
(7.2b)
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 95
, . -+ 2.r (7.2c)
fO -.21 (7.2d)
(7.2e)
(7.2f)
(7.3a)
(7.3b)
(7.3c)
A similar equality holds for the decay of the ns which is the eighth
component of an octet and depends upon the linear combination
For the decay of a flavour singlet meson, the two models give
different results:
(7.6)
M(f-)/M(AZ)/M(f') :: (7.8)
Since the decay rates are proportional to the squares of the matrix
elements, the ratio of the fO to A2 decay rates is predicted to be
9 in the Han-Narnbu model in comparison with 25/9 in the fractional-
ly charged model. Furthermore, the f' decay rate is predicted to
be larger than the A2 decay rate by a factor of two in the Han-
Narnbu model and lower by a factor of 2/9 in the fractionally
charged model. These appear to be large observable effects.
(7.9)
The relations (7.10) are derived under the assumption that the
fO and A2 are degenerate and have the same width. Calculations of
the charged and neutral kaon pair mass spectra show that the quali-
tative features of Eq. (7.10) remain when the masses and widths of
the physical particles are introduced and each resonance decay is
parametrized by a Breit-Wigner curve. In addition, a strong inter-
ference effect appears in the region between the ff and the A2 from
the overlapping of the tails of the resonances. With the Han-Nambu
model these interference effects should be somewhat different, and
might be used to distinguish between the two models.
The effects in Eqs. (7.8) and (7.10) are so large that they
may still be observable even with an appreciable reduction from the
propagators of the colour octet states. The parameter K will have
a value (m1/ms)2, where m1 and ms are the masses of the colour sin-
glet and colour octet intermediate states which are dominant in the
transitions. If K is between 10- 1 and 10- 2 , there may still be a
possibility of observing these effects. For example, if K = 2%,
there will be an 8% increase in the ratio of the two-photon decay
widths of the ff and A2, a 3% increase in the ratio of the widths
of the f and A2, and a 3% decrease in the production ratio of
charged to neutral kaon pairs over the predictions of the fraction-
ally charged quark model. Thus even if the effects are small, they
appear as uniquely related discrepancies from the predictions of the
fractionally charged model in three different ways.
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 101
Once states which are not colour singlets are produced opera-
tors which are not colour singlets become observable. However, the
parton sum rules do not necessarily hold immediately in this new
domain with the integral quark charges of the Han-Nambu model. The
basic incompatibility between naive parton models and non-Abelian
internal symmetry must be carefully considered before drawing con-
clusions. We now spell out this incompatibility explicitly and
show the necessary conditions for validity of the naive parton mo-
del.
<f IT/ ~) = 9R <tlARI i> + ~W <f'A W 'i>"'98 <fl AS' i. > (7.11)
where ~,~ and AB are reduced transition amplitudes for the red,
white and blue quark transitions with the quark charge factored out.
The values of these reduced matrix elements for different colours
are related by the colour symmetry and depend on the colour quantum
numbers of the states Ii> and If>.
For the case where both Ii> and If> are colour singlets
(7.l2a)
The contributions of the three diagrams of Fig. 7.1 are all equal
and have the same phase and the invariant amplitude A is defined
for convenience with the normalization indicated.
(7.12b)
(7.l3c)
(7.15)
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 105
Equation (7.15) shows that the naive parton model which ne-
glects coherence between the three diagrams of Fig. 7.1 is valid
only when there is a definite relation between the total cross-
sections for producing colour singlet and colour octet final states.
The exact value of the ratio of octet to singlet production depends
upon the values of the coupling constants, but ~s always of order
unity. For the Han-Nambu model this ratio can be seen to be ex-
actly unity by noting that the expressions (7.l4a) and (7.l4d) be-
come equal when the Han-Nambu coupling constants and the condition
(7.15) are used. Thus if the Han-Nambu model is correct, the naive
parton model predictions become valid only when the total cross-
sections for colour singlet and colour octet production become equal.
jugate of the diagram of Fig. 7.lb or Fig. 7.lc. When this is ex-
pressed as a diagram for elastic Compton scattering it shows an in-
termediate state undergoing a colour change. Even though the con-
ditions of the naive quark model are assumed to hold and the same
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 107
quark which has absorbed the intial photon emits the final photon,
the colour of this quark can change during the intermediate state
as a result of the colour oscillations mentioned above. These os-
cillations can be studied in detail by examining the properties of
the propagator.
(7.l6a)
(7.l6b)
(7.l6c)
The colour eigenstates [f 1 > and [fs> have different energies El and
Es because of the energy required for excitation of colour octet
states. Thus if the state [R> is created at a time t = 0, the rela-
tive phase of the components [f 1 > and [fs> change with time and in-
troduce admixtures of the other states. For example,
(7.l7a)
(7.l7b)
8.1 Introduction
(8.la)
(8.lb)
For the K*n and pK decay modes the branching ratio is unity
in the SU(3) limit except for differences in kinematic (phase
space) factors for the two final states. However, because the
two octets have opposite charge conjugation behaviour, the Al oc-
tet decay is described with F coupling and the B octet decay with
D coupling. The relative phases of the Kp and K*n decay amplitudes
are thus opposite for the two cases
(8.2a)
(8.2b)
the decay amplitudes for the mixed states (8.1) are then
<K*1C I Q .. '> = (.OS 8- < K-1t I (;lAo> + Si.flS< Klflt I QB> (8.3a)
Equations (8.3) show that for any mixing with a real phase,
the effect for one eigenstate is to enhance the K*n decay mode and
112 H. LIPKIN
and suppress the K , and vice versa for the orthogonal eigenstate.
For = 45 0 , we obtain
I(KeIG.f>I2. = I<KTC1Gz.>I2.
1(l(iJ1t IG .. >I2. I<I IG2.,>f2. - (8.4a)
of >
I... I ( K-lrl GA 12.
(8.4b)
4 - I ( K-1t 1Q S >I ~
we s ti 11 have
(8.Sa)
(8.Sb)
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 113
the decay, s wave and d wave, and the result is very sensitive to
the relative amplitudes and phases of the sand d waves. In par-
ticular, for the ratio of s to d wave amplitudes predicted by the
naive SU(6)w quark model, the transitions (8.5) vanish and cannot
produce mixing, because the QA is coupled only to vector meson
states with transverse polarization and the QB is coupled only to
longitudinally polarized states 70 ) . For this reason, the mechanism
(8.5) for mixing was dropped.
Now that the SU(6)W predictions are known not to agree with
.
exper~ment
73) ,part~cularly
. . I I I
~n the c ose y re ate
d po I ar~zat~on
. .
predictions for B and Al decays, and the experimental data are con-
sistent with pure s wave for the Q decays, the mixing mechanism
(8.5) should, perhaps, again be considered. However, a more real-
istic calculation would consider the coupled channels K*TI and Kp
through the resonance region, with phase space factors changing
within the resonances because of the proximity to threshold.
114 H. LIPKIN
The relevant sum rules are the charge exchange sum rule (CHEX)
These sum rules hold for any meson nonet and do not make any
assumption about the mixing angle, except for the conventional
description of the nand n' as two orthogonal linear combinations
of pure SU(3) singlet and octet states defined in terms of a single
mixing angle. For the case of ideal mixing, as in the vector me-
sons, the two sum rules each split into two equaltities, CHEX be-
comes:
(8.7a)
which is just the A.. Z rule, and substituting (8.7a) into (8.6a)
gives
CTln-.p ~ eO~)+CS"(Tt-p~W't1.) =
: er( K+"., + J(1tp)+CS"'CK-p ~k'''0IP1.) (8.7b)
116 H. LIPKIN
(8.8a)
(8.8b)
(8.9b)
All the vector meson relations (8.7) and (8.8) are in excel-
lent agreement with experiment. However, the pseudoscalar meson
relations (8.6) and (8.7b) are in strong disagreement. The rela-
tion (8.7a) agrees with experiment if the n is assumed to be pure
octet. This suggests that the conventional picture in which there
is small mixing may be valid for the n, but that something is
wrong with the nt, and it is wrong in the direction that the n'
has an inert piece in the wave function which does not contribute
to the sum rules (8.6) and (8.7).
0'"( A + B ~ ~ I + X )
(8.10)
cr(A+B~1tf + X)
(8.11)
The very ,
prec~se ,
exper~mental d ata 79) now ava~'I a bl e on '
p~on,
(9.la)
(9.lb)
201~----~----------------~r-------~------~
15 f(j(~P)
~~~~---
~
~
.c
(pK)
E
10
b
O~ __~~____~~______~~____~~____~
6 200
Plab (GeV/c)
(9.2a)
(9.2b)
(9.3a)
(9.4)
WHY IS THERE CHARM, STRANGENESS, AND COLOUR? 123
(9.5a)
.,
CTR (Hp) = (N~ +2.N;)( P&llof I (9.5c)
The formula (9.4) predicts that plots of Gtot (Hp) (P lab /20)
o
(E+O)
vs Plab should show straight lines for all cross-sections and
linear combinations of cross-sections which have no Regge contri-
bution. This is strikingly verified in Figs. 9.2a and 9.2b, which
show straight lines for Gt (K+p), 0' (pp), and for the linear
ot tot
combinations (9.1) and (9.2). A straight line is not obtained for
Gtot(n-p), which has a Regge component.
124 H. LIPKIN
(9.6)
N
0....... ~O b.
u b. b. b.
.......
0 b.
> Q)
b.
..
t.:>
.0
30 0
b.b.
oo~ b.
~
II
Ob.b.
E ~ /:,b. i
N II !i!
allo ,...
....
0_ b.b.
II
o 20 0 ,,_ill !1
,,-
N
_II-
.............
.0
0 "
(1.-
ti llV .....
-
~
10 II
" ~~
0
b- !j!!i! ljl ~ljl_ II ~II-"e .. It III tI~ ti ti
0
0 2 3 4 5 6
0.33 0.33
(~ab ) (G eV Ie)
/:, tot(n-p)
0 (2/3)Otot(pp)
tot(n-p) - R(n-p)
+
'- tot(K p)
x (<jJp)
1/ ol(pK)
0 /:'(HB)
+ /:, (nK)
t..)
0.
Table 9.1
Theoretical predictions and experimental data for tot (pp) and p(pp)
I
1491 52.9 43.5 42.5 42.4 0.4 0.079 0.078 0.010 I
i
300 I I I
'"
0_
()
.......
>.,
250 f -
(!)
I
200 - -
.a
E I
'"0_ 1501- -
0
C\I
..............
.a 1001- -
c
0.-
'0 50 1-/ / -
b-
I I I
Oo 10 20 30
0.33 0.33
(Plab ) (GeV Ie)
(P )0.33
Fig. 9.2b (Jtot(pp) x (P lab /20)2 plotted against lab
REFERENCES
7) G. 't Hooft, Nuclear Phys. B33 (1971 173; B35 (1971) 167.
21) Mary K. Gaillard, B.W. Lee and J.L. Rosner, Reviews of Modern
Physics, 47 (1975) 277.
24) The rule forbidding ~ + p~ and other stuff has been credited
to Okubo, Zweig, Iizuka and others in~rious combinations.
To avoid arguments about credit, one can refer to the A Z
132 H. LIPKIN
rule and allow the reader to insert the names of all de-
sired friends from Alexander to Zweig. For a general re-
view, see, H.J. Lipkin, in New Fields in Hadron Physics,
Proc. XI Rencontre de Moriond, ed. J. Tran Thanh Van,
CNRS (1976), p. 327 and Lectures at the Erice Summer School
(1976), Fermi1ab preprint Conf-76/98-THY to be published
in the Erice Proceedings.
26) B.W. Lee, in Proc. XVI Int. Conf. on High Energy Physics,
Chicago-Batavia, Ed. J.D. Jackson and A. Roberts (1972),
Vol. 4, p. 249.
27) H. Georgi and S.L. G1ashow, Phys. Rev. Letters 28 (1972) 1494,
Phys. Rev. D7 (1973) 561.
28) J.D. Bjorken in Proc. XVI Int. Conf. on High Energy Physics,
Chicago-Batavia, ed. J.D. Jackson and A. Roberts (1972)
Vol. 2, p. 299.
47) J.C. Pati and C.H. Woo, Phys. Rev. D3 (1971) 2920.
R.L. Kingsley, Phys. Letters 40B (1972) 387.
See also Ref. 8, p. 94.
51) S. Okubo, Phys. Rev. D13 (1976) 1994 and D14 (1976) 108.
55) R.L. Jaffe, Phys. Rev. D15 (1977) 267 and 281.
61) R.D. Carlitz, S.D. Ellis and R. Savit, Phys. Letters 68B (1977)
443. See also N. Isgur and G. Karl, Phys. Letters, in
press.
66) H.J. Lipkin, in High Energy Physics, Proc. 1975 EPS Conf.
Palermo, ed. A. Zichichi, Editrice Compositori Bologna
(1976) p. 609.
D. Faiman, H.J. Lipkin and H.R. Rubinstein, Phys. Letters
59B (1975) 269.
67) M.Y. Han and Y. Nambu, Phys. Rev. 139 (1965) Bl006.
73) J.L. Rosner, in Proc. XVII Int. Conf. on High Energy Physics,
London (1974), ed. J.R. Smith, p. 11-171, Phys. Reports
llC (1974) 190.
76) E.M. Levin and L.L. Frankfurt, Zh. Eksp. i. Theor. Fiz-Pis'ma
Redakt 2 (1965) 105 (JETP Letters 2 (1965) 65).
77) S. Okubo, Phys. Rev. D13 (1976) 1994 and D14 (1976) 108.
80) H.J. Lipkin, Phys. Rev. Dll (1975) 1827. There are some
errors in the values of the parameters quoted arising from
a confusion between several notations. The correct values
of the parameters actually used in this reference are
given with Eqs. (6.4) and (6.5) of the present paper.
See also, Nuclear Phys. B78 (1974) 381; Phys. Letters 56B
(1975) 76; High Energy Physics, Proc. of (1975) EPS Conf.,
Palermo, ed. A. Zichichi, Editrice Compositori Bologna
(1976) p. 1000, and FERMILAB-Pub-77/49-THY.
DISCUSSION No.1
SCHNEIDER:
I have a question about your discussion of Han-Nambu
quarks, and the difficulty of distinguishing experiment-
ally between the Han-Nambu integer charge assignment and
the Gell-Mann-Zweig fractional charge assignments.
Could you explain this in more detail?
LIPKIN:
We are again running into the problem of reconciling
Quantum Mechanics with the parton model. Parton distri-
butions contain only intensity information, not phases,
and you will quickly run into trouble with QM if you
forget this fact. Let me give you an example. Consider
the case of the isoscalar EM current coupling to a pion.
This current has odd G-parity, so the final state X in
1{ lc:o...n:. ... ')( must have even G-pari ty. Now observe how
this conservation law is "violated" in the naive parton
picture. In that picture a quark absorbs the photon
with a probability given by F(x).
Probability N 'f()t) .
137
138 DISCUSSION
COLEMAN:
SCHNEIDER:
LIPKIN:
JEZABEK:
R = crle+e,---, h~("on$) =
(3" (ete: - P+\L" )
LIPKIN:
CHEN:
LIPKIN:
MIETTINEN:
y;Y /h
~x
(el)
~. X
(I
DISCUSSION 141
WEILER:
COLEMAN:
NILSSON:
LIPKIN:
DINE:
LIPKIN:
'c.
itative things to say on this question.
As for ~he it has many decay channels open with
small branch~ng rat~os, so it would not seem surprising
if it has not been seen yet. However, the Se interpret-
ation of the y events at DESY is an idea which has
occurred to me very recently, and there could be data
available which would contradict this interpretation. I
have not had time to look into this yet. Perhaps I will
discuss this in more detail in my next lecture.
DINE:
Would the standard charmonium picture be in trouble
if your interpretation is correct?
LIPKIN:
No. In fact, It Hooft would like very much for the
'1., to be higher.
CELMASTER:
The processes ~ -''lY ) i' ~ yt'y I and ~ -. 'X(2..t.) "
is the 'c .
have been seen at DESY. You have suggested that X(2.8)
Judging from the fact that the first two
processes occur, we would guess that there is some mixing
of co pairs into the light mesons. Hence one might guess
that one could see also ~- b"f Why is this process
not seen?
LIPKIN:
The answer to your question is that if &c is a mole-
cule, it is difficult to say how much mixing will occur.
SANDA:
I have a comment about the possibility that X(2.8)
is a Sc' You mentioned in your lecture that -.!J'.... p Sc.
exists. Then ""' .. "( Sc also exists, and DESY should have
a good upper limit on this branching ratio since they
are very sensitive to a monoenergetic photon.*
BALDINI-CELIO:
If "l' -'> 1'\ '( and '\J - rt'Y are larger than '-Y~ 1(0"(
because the former processes pass through a two-gluon
intermediate state and the latter through a three-gluon
intermediate state, as you have claimed, then why is
i"-Il "" so large?
LIPKIN:
You can either include the cc ~qq as part of the '1. wave
function and say that it is mixed, or you can say that
the transition is a two-step process ""/~'tY'/C-'t'1.
The result is the same to first order in the OZI-
violating diagram.
BALDINI-CELIO:
LIPKIN:
DEO:
LIPKIN:
ft+----r-- A+ .+ 6+
f)A,. p,A a
A
-0
_---'L..-- -+
~- tf
!:::.
M1 M20
COLEMAN :
I can write
LIPKIN:
The idea is that if a baryon-anti-baryon annihilate,
you can get two quarks coupled as a vector and two anti-
quarks coupled as a vector. The question is: can it
decay to two scalars?
-=--
v
" 1/ .=0
f. S
-
0 ~
0 - II +
#~ 0 o=-e s
0=0
v
JEZABEK:
LIPKIN:
With Han-Nambu quarks, the electromagnetic current
cannot be a color singlet so color cannot be an exact
symmetry. There must be high mass states which are not
color singlets.
DISCUSSION 147
JEZABEK:
LIPKIN:
MESS:
LIPKIN:
TOWNSEND:
LIPKIN:
MORINESCU:
COLEMAN:
KOCKER:
MIETTINEN:
LIPKIN:
MIETTINEN:
LIPKIN:
MIETTINEN:
DINE:
LIPKIN:
COLEMAN:
BERLAD:
& :;? M
~ ~M
forbidden by the Zweig rule?
LIPKIN:
is allowed, but
is forbidden.
t ___
~ ---..
One can define this selection rule ~n either of two ways:
1) hairpin diagrams (disconnected qq pairs of the same
flavor) are suppressed; 2) diagrams with disconnected
parts are suppressed. For 3-quarks, both give the same
Zweig rule, but with 4-quarks, there are ambiguities.
For instance
::::>,,1/
.
is forbidden by both definitions. The graph
,\,1
--~
---~-
---
is prohibited by Rosner, but allowed by Jaffe. Also
V~
~Tr
presumably is double Zweig forbidden. My conclusion is
that there is no dynamical understanding of the Zweig
rule, so the generalization to the 4-quark case is not
unique, and may be impossible.
BIZZARRI:
LIPKIN:
Q
where the two Q's are outside the range of the potential,
but the two QQ pairs are inside. Although this might
work, it is unrealistic to assume that the potential has
sufficiently sharp sides. In the more realistic cases
of smooth sided potentials, it does not work.
COLEI1AN:
LIPKIN:
This is true. The four quark configuration is not
strictly a molecule but more like a microcrystal, so
you might still expect binding. But even though the
four quarks are not arranged into two color singlet
QQ pairs, it still does not work.
BERLAD:
Can you show us the five parameters fit to the total
cross section you showed us this morning and explain the
two new terms? Could you also explain how you obtain the
real part?
LIPKIN:
The fit is to
where
Cf = b.S Mb
C2. = 2..'2. mb
c~ = 1.15 mb
t = C.I:' ,) S :.0.2.
DISCUSSION 153
BERLAD:
LIPKIN:
Of course. If you go to high enough energies this
model won't work. But it is the simplest thing to do;
anything better than a power would involve more parameters
Getting the real part is trivial, because if you have
an analytic expression involving powers you get an express
ion for the real part by putting in the standard phases.
MARTIN:
As long as you use this expression far from the regio
where it breaks down, there is no problem with getting the
real part.
BERLAD:
But if you use this expression you may need a second
subtraction, whereas the usual expression has one sub-
traction.
MARTIN:
Yes, but for an order of magnitude estimate it is OK.
WEILER:
Two conclusions from this morning are: first, the ~I
is heavy; second, the ~ has components other than SU(3)
octet and singlet. Can both of these problems be approach
ed by mixing in some cc? Does this conflict with known
-:r/~ branching ratios?
LIPKIN:
The problem of the ~' being too heavy is not solved
by mixing in something heavier. The wave function for the
lowest state having a given set of quantum numbers can onl
be pushed down (variational principle). From the point of
view of perturbation theory, the off diagonal piece
dominates over the additional mass term.
DISCUSSION 155
DINE:
LIPKIN:
HINCHLIFFE:
LIPKIN:
BALDINI:
LIPKIN:
MARTIN:
I have several comments on the pp cross sections. I
noticed (as did Arthur Wightman) that the highest energy
points of the CERN-Rome experiment sit below their best
fit. According to Ugo Amaldi this reflects an energy-
independent uncertainty in their overall normalization.
This normalization is left free, within reasonable limits,
in their fit.
DISCUSSION 157
LIPKIN:
R. Bizzarri
1. INTRODUCTION
P
C
(_l)L+l
(-1) L+S
Le., J L for S
-+ -+ -+
J L-l, L, L + 1 for S 1 .
J L + S
159
160 R. BIZZARRI
Table 1
J PC
iI L S
--
I = 1 1/2 0 0
0 1 1 p K* (892) w
0 0-+ K (958)
11 11 11'
0++
P (970)
..
K
... E: ~:..(??.3).
+-
0 1 B (1235) QB - -
--
..
2 1 3 g (1680) (1800) w (1675) -
-- ~........
2 - - - -
--
1
p'.........
(1600) - - -
+-
0 2 (1640) L (1770) - -
A3......... ........
DO MESONS FILL SU(3) NONETS? 161
Table 2
TI P (A l ) 1080 K 730
Table with Table 2, taken from the lectures by S.M. Berman at the
1964 School of Erice (with present day names added) which showed
the observed "meson resonances which at the present time are not
fitted into some SU(3) multiplet structure"l). Of these states the
++
A2 is a well-established member of the 2 nonet, the B is a well-
established 1+- state but still without a nonet. The existence of
an S* pole is still questioned, strong doubts exist on the resonant
nature of the Al , the quantum numbers of the E meson are not estab-
lished. As for the S-wave TITI and TIK resonances (<: and K) their
presumed mass has been steadily moving upwards and only recently
began to stabilize around 1300 MeV. The fact that 13 years have
not been sufficient to clarify the situation is somewhat worrying
and might be taken as an indication that something does not work ~n
2. The 0+ Nonet
2.1 I = 1
25
>
GI
(!) 20
('II
o
o
......
15
en
~
Z
UJ
>
UJ 10
o
UJ
~
X
(!)
UJ
~
0 .6 0.8
10
~
(!)
N
o
..g
en 5
~
z
UJ
>
UJ
o.~6----~-----A~~--L-----~~WC----~
0.8 1.0 1.2 1,4 1.6
M (K --;KO) , GeV
above KK threshold
below KK threshold
where qn is the 11 momentum in the llTI system, ~ and g.!] may be con-
sidered as squares of the coupling constants to the KK and llTI sys-
tems, fo is a measure of the coupling to the initial state, it may
be considered constant ~n a production experiment.
2. \' n\
c.. m~ 0 1t'1
., bl
"' ~ _ ...v ""R::' eoo ".-"I
0 :;: 82 M.V r 0 ;; 3C() III_V
C : 91 J'b/GoV C = ~,.IjIG .v
O ~~~~~~~~~~LL~~
80
920 1040
MASS OF DECAYING SYSTEM (MeV)
Fig. 2
-
KK and TIn mass distributions from Ref. 2. The
curves represent the theoretical distributions
obtained from a two-channel resonance with
parameter values given in the figures.
166 R. BIZZARRI
2.2 I =0
The I = 0 members of this nonet are expected to be strongly
coupled to the TITI and KK systems. The TITI interaction is not direc-
tly accessible to experiment but it can be studied by interaction
of pions on virtual pions with reactions like
-t A++ (2)
"It. p~ u n+7t-
G'2. -t:
=
47t.
n_ __n
----Y:-=--=--=---n
n: I
I
N
Fig. 3 The one-pion exchange contribution
to dipion production.
3roOO ~----~t~----------------'
++
.......
>- +
II>
+ +
-
~
o +
N
.' + +
en
I- t .
+
z
UJ t f
>
UJ + +
'.J
O ~~~~--~----~----~-------J
0.2 0.6 1.0 1.4 1.8 21
+ - - +
Fig. 4 TI TImass spectrum for events TI p ~ TI TI n
with It I < 0.15 GeV/c. Histogram: ob-
served events. Dots: weighted observed
events. Points with error bars: pro-
duced events.
168 R. BIZZARRI
~O'r-----------------------------------~
1.0
U5 1):
a
180"
150
,
{j 0
120
9(t
SOo
30
1.0
Q5 rf,
a
10 ~
0 ---
170
1.0
0.5 ,
1)'
-tt
m
s"*'
~ 'l000 MeV rS = 60+200 MeV
- S* S*
and its coupling to the KK system must be very large: g-- > g--
KK TITI
Therefore this experimentally narrow effect is again associated
to a resonance which can be considered "wide" ir: the sense that its
-
observed width is limited by the nearness of the KK threshold and
not due to a small value of the coupling constants.
~0
o
= bO0 _ \ b: ') '0,"'"
DO MESONS FILL SU(3) NONETS? 173
3600
3300
3000
2400
1.0
0.5 1JO
D
o
6
Fig. 8 ng
The I = 0 s-wave TITI phase-shift 8~ and inelasticity
as obtained from energy independent (points) ano energy
dependent (curves) fits.
174 R. BIZZARRI
2.3 I = !z
The I = !z S = l members of this nonet are expected to couple
to the S-wave Kn system which can be studied from the reactions
K P -+ K n+ n (4)
K p -+ K n- 6++ (5)
2.4 Conclusion
The most recent data suggest the existence of a full 0++ nonet
made of resonances with a very large width.
DO MESONS FILL SU(3) NONETS? 175
150
d~
(degrees)
120
90
60
0.2 1.6
(8,3 03
SOLUTION A SOLUTION 8 SOLUTION c SOLUTION 0
'3
1.1 1,1 1,1
( )
S' 1.5 S'
1.5
1.7 1.7
1.7 ,5
M,.5 /
~~'3 P,,-~
1.7
1,5 1,3
Fig. 10 Argand diagrams for the four solutions for KTI partial waves.
DO MESONS FILL SU(3) NONETS? 177
1300 MeV and m ~ 1400 MeV but the uncertainties are large and
K
m > m cannot be excluded. In this respect it should also be
E K
noted that a clear and practical definition of the resonance mass
in the case of wide resonances in the presence of large background
is lacking. Certainly the mixing for this nonet is far from ideal
and the mixing angle 6 should be quite large.
3. 1+ Nonets
3.1 I = 1
TI P -+ TI - TI - TI +P (6)
--z
N
(I) 250 a) 0.04<ltI<0.17 !GeV/d
~
~
:>
~
2.5
b) 0.17<ItI<o.33 IGeV/d
:>
II)
~
C
--
N
500
~
z
~
:>
~
'\, 1150 HeV and width '\, 300 HeV. To deduce the properties of this
3TI system a partial wave analysis is performed. The cross-section
of reaction (6) is written
(7)
are the production variables for the 3TI system, its decay variables
being
Pab is the density matrix for the production of a 3TI state. Such
a state is assumed to have spin-parity and polarization JPM and to
decay via an I-wave into a TITI system of spin S. Therefore the
meaning of the indices is a = JP MIS.
where m 3 rd component of 1
)l 3 rd component of S
8 1 l PI angles and momentum of (TI + TI -2 ) in the 3TI c.m.s.
8 1 l ql angles and momentum of TI +.1n TI.+-
TI2 c.m.s.
with
q 2 'h \
is (S.1') = Vs (-!) ~s
'10 .fS1
and the values of mS and fS being
S = 0 (E) M
0
760 f
0
400
1 (p) M 760 fl 138
2 (f) M 1265 f2 154
The data fitted with formula (7) and the values of Pab deduced
as functions of the 3rr mass. The P aa with equal indices give the
intensity of production of state a, the argument of Pab gives the
relative phase between states a and b.
The method used to describe the 3rr system in terms of the final
state rrrr interaction can be and has been criticized on theoretical
grounds, in fact the amplitudes used violate the requirements of
181
DO MESONS FILL SU(3) NONETS?
0.17<ltI<Q.33 (GeI/C)2
400
i
~300
en
t-
z
~200
100
1.0 1.1
100
!
200)
fHf~
" S (Qnl
500
ff ~ "0 (QnJ
...
f
~
-
~
U")
N !f 0
! ! 2-P
en OXI
t
~
zu.J (pnJ
f wiltffl~
:>
u.J
:>
II>
,! ~
.++'+
! !
,. P (en)
vsl"P (Qnl
!t
S
u.J
~ ~~{I}iiOQo?Oc.:;\#? ,~+!
~
~
0
-soo
141 ~~~~~~li~\sl'p (enl
u.J
u
zu.J
a:
.... -me ~ . 02
u.J
a: Y ,~~oQ 500 o-s (en)
u.J
~
-1 Q~ vsO-S(enJ
~
3.2 I =1
The strange mesons are not eigenstates of C nor of the G-parity
operator. We cannot therefore distinguish between the two states
QA and QB belonging to the two nonets with J PC = 1++ and 1+-. If
however there are two nonets there should also exist two physical
mesons, both with JP = 1+, which can eventually be a mixing of the
184 R. BIZZARRI
01..lL.JL.:....L----"':.-.c~&J
K- It It-" 13 GeV
- 1'< 0.3 GeI/
--;:;---- 1500 a)
--"I:';
c.::J
"":;:;-
::I.
-.." E
.."
150
-:;---- C) d)
;::.
~
..,.
........::...
';,.g
"1==
... 360
'"'"g.
"E 180 ~~
..,."'-
0
I I
0
120 e) fl
..-;:;-- rl+KIt
~
--'fc
~
.D
80
::I.
40
-.." E
.."
0
180
...!!:!
-;;;
en
:s'" Of-- . \~
t
- tf_+~
..,.
r r I
- 180
1.0 1.2 \.4 1.6 1.0 1.2 1.4 1.6
m(K ltltl (GeV) m(Kn: It ) (GeV)
600 r------~
a) b)
>
'"
t!:)
o "C
-
';:,E
"C
90 1-
01-
- 90 -
x
+
- 100~~'~1~~L-'~~
240 c) d)
...
~
>
'"
t!:)
270
...
......
'"~
0>
180
'"
:s
90
and
In this mass region Knn enhancement has been, for a long time,
observed in non-diffractive reactions, the first observation in pp
annihilation (known as C-meson) going back to 1964 15 ). I shall
only mention the very recent results from the ACNO collaboration 16 )
from a 133 events/~b bubble chamber exposure. The reactions
- - +
K P -+;::. (KnTI)
3.3 I = 0
> 150
Q)
:2
0
N
-
"'-..
'"
c:
Q) 100
-
>
Q)
c
....
Q)
..c
E 50
::J
;;;::
1.8
Mass ( Kn n) GeV
Fig. 18 Knn mass spectrum from reactions K-p + ~- (Knn)+
with u < 1.5 (GeV/c)2. The full line curve is
the result of the fit to determine the Ql mass
and width.
DO MESONS FILL SU(3) NONETS? 191
4. Conclusions
or
+ -
e e -+ CP' -+ QK
The ma~n advantage would be to have a situation free from any dif-
fractive background. Good 4n magnetic detectors on low energy e+e-
machines would be necessary.
References
DISCUSSION No.1
HINCHLIFFE:
BIZZARRI:
PREPARATA:
LIPKIN:
195
196 DISCUSSION
ZICHICHI:
LIPKIN:
No, I can't.
DEO:
BIZZARRI:
PREPARATA:
WETZEL:
MANDELLI:
BIZZARRI:
WETZEL:
MIETTINEN:
~0
2 -
(b) '" - W .". 1r
i30
"en
~20
."-
III
>
III
10
O~~----~--~--~--~
500 1000 I~ 2000 2500
m(.".+ ".-) (MeV/CI )
198 DISCUSSION
BRANSON:
MARTIN:
This morning you mentioned that the problem of Chew-
Low extrapolation was delicate. In fact there are two
problems. One is the extrapolation to get the Pion_Pion
cross section. Another one is that of getting phase
shifts from differential cross sections and this last
process leads to non-unique answers as I explained in my
1975 Erice lecture notes. The only thing which restores
uniqueness is analyticity with respect to energy and this
has in fact been used in practice by Alan Martin and
Pennington. Otherwise you may have two sets of phase
shifts reproducing the same data as has happened in the
past to one of the CERN-Munich experiments.
BIZZARRI:
Yes. There are three relevant points: first to ex-
trapolate the cross sections, then to get a set of phase
shifts. There you have ambiguities but even when you
resolve all ambiguities, which you can by imposing anal-
yticity and constraints from other reactions, you have
to read them in terms of poles in the complex plane and
this third step has still a lot of arbitrariness.
DISCUSSION 199
DISCUSSION No.2
JENNI:
BIZZARRI:
BALDINI CELIO:
BIZZARRI:
PREPARATA:
MIETTINEN:
MIETTINEN:
BIZZARRI:
I forgot to comment on the fact that the two strange
members of the 1~~ and 1+- nonets are expected to decay
into Ie. p and "*'71 with equal probabilities. The observa t-
ion of almost pure states like ICp or IC-*TC. is evidence of
mixing. I do not see how from this mixing between two
different nonets you can deduce the properties of the A1
MIETTINEN:
BIZZARRI:
I know of the spectrometer study of the reaction
7t' -p -+ ( 7T "'lcn 0) + t'\
~yy
but I do not think they have yet published a conclusive
analysis of the Ai region. Their data will be very in-
teresting. Of course in this reaction there is also the
1=0 contribution to the 3J(. system. I do not think that
the existing evidence against non diffractive Ai prod-
ution is really so significant. The possibility you are
advocating that the peak seen in the 4 GeV/c ~-p experiment
DISCUSSION 201
MIETTINEN:
BIZZARRI:
I don't know.
ANON:
MIETTINEN:
LIPKIN:
Herwig Schopper
Deutsches Elektronen-Synchrotron DESY, Hamburg
and
II. Institut fUr Experimentalphysik der Universitat Hamburg
1. INTRODUCTION
An impressive amount of data has been accumulated over the past three
years which support strongly the idea that elementary particles are composed
of four quarks instead of three. All the predictions based on the existence of
the fourth, the charm quark, have in principle been borne out by experiments, and
from detailed investigations very interesting results could be obtained for the
strong and the weak interaction. Most of these results have been obtained with
electron-positron storage rings, and only these will be discussed in this re-
port. Additional data from hadronic or neutrino interactions are in general agree-
ment with the e+e- data and will not be discussed here. This series of lectures
will be entirely devoted to a discussion of mesons, i.e. quark - anti quark
systems. Although some indications have been found for the existence of charmed
baryons, these data are still rather scanty. In the last chapter we shall also
discuss the upsilon particle.
203
204 H.SCHOPPER
The two rings of DORIS would make it possible also to study electron-electron
and electron-proton collisions. These options have not been used so far, because
of the strong interest in e+e- physics.
The layout of DORIS and the injection scheme shown in fig. 1.1. So far DORIS
mostly has been operated at energies between 2 x 1.5 and 2 x 3 GeV. The electron and
positron currents are usually about 200 mA and the beam lifetime is varying
between 5 and 12 hours. Currents between 0.5 and 0.8 A have been achieved, but
the beam lifetime then becomes short and the momentum smearing in the beam gets
bigger than 1 MeV. In October 1977 DORIS was converted to single ring opera-
tion and energies of 2 x 4 GeV have been achieved. The limi-
tation is then given by the rf-power. It is planned to add two more PETRA
cavities in DORIS, which should make it possible to push energy up to
2 x 5 GeV. This will be tried during 1978.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 205
r:::. ~- _ _
~ Electrons
~ SYNCHROTRON
m~~'\~ , Positrons
W :Interaction Point
LlNEAR-~ \\
ACCELERATOR I
r' '
LlNEAR-
ACCELERATOR 2
600 MeV
.7
, E(GeV Ibe.3m
206 H. SCHOPPER
[nb'~ ,
AVERAGE WMiNOSf1YIWEEK
OF OORIS
f
\00 8 t---
I~' ~
1~0
~
l- e- t1 t---
o JJ1l
r
I() liJ 30 III
1976 --~ 19n--
\0
j
I() WEEK
Fig. 1.3
Fig.I.3 shows the average luminosities per week obtained during '76 and beginning
of '77. It is the average luminosity which is really important for the experi-
ments. The continuous increase of the average luminosity is mainly due to im-
provements in the stability and reliability of the machine. With a total cross
sect i on of about 40 nb several hundred events can be observed per day.
l~!f_~g~~trQ~~t~r~
r/J -3m
t -3m
, kr
SPEAR SLAC-LBL MAGNETIC DETEClOR
eo'eo- COLLISIONS
Fig, 1.4
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 207
Iron yo~.
"==01 -----tlt+--+-..,J,~
JO~lI'< 1'11<,"
l.I05crn
TtllXlOf Ace:
6,R .81 ,X"
Fig. 1.5
very good momentum analysis, which is one of the strengths of this spectrometer.
A solid angle of about 65 % of 4n is covered. Electrons and y-rays can be ob-
served by shower counters, which are placed outside the coil. Time-of-f1ight
measurements with scintillation counters make it possible to identify particles,
and in particular the separation of K and n'S turned out to be very important
for the discovery of charmed particles. Large spark chambers located outside
the iron yoke allow the detection of muons. Since the iron yoke is not very
thick, there is however some punch through from hadrons. In order to provide
cleaner muon detection for at least a limited solid angle, additional concrete
absorbers were later added on top of the detector ("muon tower") . Very
recently a lead glass wall with an active converter has been added, which has
proved to be very important for the clean detection of electrons.
Fig.1.5 shows the PLUTO spectrometer at DORIS. PLUTO also uses a longitudinal
magnetic fie1d,which, however, is produced by a superconducting coil giving a
field of about 2 tes1a. Since the diameter of the coil is only about 1 m, the
momentum resolution is lower than that of SPEAR and time-of-flight measurements
cannot be used to identify particles. On the other hand the cylindrical chambers
inside the coil cover about 86 % of 4n. The cylindrical chambers are interspersed
with two lead converters (0.4 and 1.7 radiation le~gths thick, respectively),
which allow the detection of y'S and the identification of electrons. Large
chambers outside the iron yoke are used to detect muons. Since the iron yoke
208 H. SCHOPPER
provides a hadron absorber more than 60 cm thick in all directions, very good
muon identification can be obtained. Indeed, the misidentification of a hadron
for a muon is less than about 3 %. Also electrons can be identified quite re-
liably, again with a misinterpretation probability of only a few percent. These
properties have been very important in verifying the existence of a heavy
1epton.
IRON ABSORBER
~IRON
Q
ABSORBER SHOWER
COUNTER
analysis covers about 70 % of 4~. This central detector allows the determination
of the direction of y's and charged particles with an.accuracy of about 20 , which
has been very essential in some of the experiments.
!~!~ __ I~Q~~_Qf_~~Q~r!~~~~~
If an electron and a positron annihilate, a virtual photon is produced
which then decays either again into a lepton pair or into hadrons (fig. 1.7).
The total final state must have the quan-
tum numbers of the photon JP = 1-. Most
a) eTtct
inclusive spectra of the experiments which have been per-
formed and which will be described later
fall into the following three classes:
1. Non-resonant hadron production (fig. 1.7a).
Here one might ask questions like what is
the multiplicity of produced hadrons, how
does the total cross section as a func-
tion of energy (excluding resonances) vary,
what are the inclusive spectra of diffe-
rent kinds of particles like, etc.
2. If the energy of the virtual photon
coincides with the rest energy of a
vector particle having the quantum num-
bers of the photon, the hadron production
Fig. 1.7 will be enhanced. Measuring the total
cross section or special channels one will notice a resonance-like structure
if the data are plotted as a function of the e+e-energy (fig. 1.7b). This is a
very powerful method to detect new particles; however, only particles with
the quantum numbers of the photon can be discovered in this way.
3. Other states can be found in the following way: assume that the virtual
photon decays into two resonances (fig. 1.7c) which in turn decay into
other hadrons. If the momenta of the final particles are measured, then
one can combine two or more particles and calculate from the relativistic kinema-
210 H.SCHOPPER
tics the invariant mass of the object from whose decay they originate. Since
one does not know a priori which particles come from the decay of the resonance,
many wrong combinations of course are taken, which produce a smooth background.
The right combinations of particles,on the other hand, produce a sharp peak on
top of this background. In this way states with other quantum numbers can be
found; of course it is necessary to measure ~ decay particles in order to
determine the invariant mass.
4. The kinematics allow us to calculate also the mass of the system recoiling
against the invariant mass determined as described. If the recoiling state also
contains well defined resonances, peaks in the distribution of the recoil mass
will be found. Since the recoil mass can be calculated from the momentum of
the first resonance and the initial state, it is not necessary to observe the de-
cay products of the recoiling resonance.
The purpose of this review is to summarize the experimental facts and not
to discuss various theories. To interpret the data, phenomenological models
will be used which have been developed over the past yeal's and which are based
on simple concepts. Most of these models lack a rigorous foundation, and
hence they are criticized by many theorists. On the other hand, these models
have exhibited a surprising and extremely successful predictive power, and there-
fore the usefulness of such models is beyond any doubt. Before discussing these
models, it might be useful to recall some generally accepted ideas.
1~f!_Q~2r~_~Q~~1
The following discussions will be based on a SU(4) xSU(3) quark model im-
plying that quarks have 4 flavours (u,d,s,c) and 3 colours (blue, red, yellow).
One of the main questions will be to clarify if the experimental results are in
agreement with the existence of a charm quark, which has been requested to
restore quark-lepton symmetrylo6) and to explain the absence of neutral currents
in K decay107).
flavour Q 13 S C Y
and baryon number B = 1/3, spin = 1/2 for all quarks with Q = 13 + i (B+S+C).
The SU(4) flavour symmetry is broken by the different quark masses.
8 +1+3+ 3+ 1
with C 0, 0, -1, +1, O.
The octet and singlet with charm charge C = 0 correspond to the old SU(3)
symmetry. The two triplets require the existence of 3 particle states with
charm C = +1 and -1, respectively. Finally the last singlet is associated
with a state ce, the two charm charges cancel and hence C = (1 ("hidden charm").
!;~~_~~rQ~Qg~~~~!~~_g~~
QED can be generalized to a gauge field theory with local SU(3) colour
symmetry 1). The interaction betvieen quarks is mediated by gluons in analogy to
photons in the electromagnetic case.
212 H. SCHOPPER
~
{ -------g
e vadfilIT'
e
>------y
UCD o..ED
as(EO}
a (E) =--~25--~(-E-}--ln--(-E/-E~}2~ ( 1.1)
s 1 + T2iT as 0 0
(1. 2)
+ L( r)
The first 1ine is, of course, the free particle term, where ml and PI stand for
the masses and momenta of the 2 quarks. The second line is associated to the
short range interaction whose radial and spin dependence expressed by S12 are
assumed to be the same for the Coulomb force and the strong interaction. The
only difference is the replacement of the fine structure constant a by as of
(1.1) and of the electric charges Ql Q2 by a colour factor k which is
k = - 4/3 for a qq system and k = - 2/3 for a three quark state.
The third line of (1.2) stands for the long range binding potential. A
preferred guess for L(r) i s 5)
L(r) =a r (1. 3)
the properties under Lorentz transformations are under question. A first guess
might be that L(r) transforms like a Lorentz vector but scalar contributions are
possible. Various models differ by the assumptions made for L(r). Recently also
logarithmic potentials have been considered l34 )135).
(1.4 )
1 1........
- - {-2 (r x PI)
2r 3 m1
The second term in the first line describes the spin-spin interaction
giving rise to the hyperfine splitting, the third term gives the tensor
and the second 1i ne the Lx S coupl i ng. The thi rd 1i ne ori gi nates from re:-
lativistic corrections of order (v/c)2. The last line has no classical
analog, its origin is the reduction of the relativistic ai-matrices to
the Pauli 0i spinors.
!;g~g __ ~2r~_~2E~i~~i~~~~_~2~~1~
a) General potential V(r)
Some authors 7 ,8)consider a more general r-dependence of the potential
V(r) than the Coulomb like l/r. In this case the gluon propagator has
to be modified
)J 1
Yl)J Y2 . 17 -- Yl)J Y2 )J v(k2)
where v(k 2) is the Fourier transform of V(r).
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 215
For simplicity S12 is given only for the case of equal masses of the two
interacting quarks:
4 = _1_ (d 2V
- "3 Ct s S123m2
_ 1.dV ) [->
-> -
s 1 ->s2 - 3 (s1 r ) (~2;) ]
dr r dr
2 ->->
+- (s1 s 2) ,,2 V( r) (1. 5)
3m 2
-> ->
+ _3_ (1 dV (
2m2
r dr r x -p (s1 + ~)
For V(r) = - Cts/r equ. (1.5) is transformed into (1.4). If the "naive"
potential V(r) = - Cts'r + ar is chosen then the confining potential contributes
automatically to spin dependent effects.
The level ordering for quite general classes of potentials has been studied
by Grosse and ~'artin8b).
( 1.6)
where aeis the anomalous moment. As a result expression (1.5) has to be modi-
fied in the following way:
2( 1+~2
+ 3m + (spin-spin term) ( 1.8)
3 24':.
+ zmz (1 +~) + (LS-term)
+ unchanged term.
e) Annihilation graphs
In the charmonium system virtual transition cc ~ y ~ qi qi are possible,
where qi are other than c-quarks. These virtual transitions effect the bound
cc state and lead to non-negligible corrections 14 ).
The influence of the improvements of the standard model will be discussed
below for particular measurable quantities like level splittings, transition
probabilities, etc.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 217
These models yield level schemes with quantum numbers which differ from
the standard model. However, it seems that they did not really help to solve
some of the difficulties which will be discussed below and hence these models
will not be discussed further.
218 H. SCHOPPER
2. MASS SPECTRA
( 2.1)
Since a priori it is not known which particles in the final state ori-
ginate from one resonance, one has to try all possible combinations.
The wrong combinations and pure phase space decays produce a slowly
varying background in the distribution of the invariant mass. Ex-
perimentally mass resolutions of the order of 20 MeV can be obtained.
Sometimes the identification of the particles in the final state is
not or only partially possible (e.g. ~-K separation). In such a
case the wrong assignment of a particle mass to a certain track
leads to "kinematic reflections" in the distribution of invariant
masses.
Q= ~10 - m1 - m - p2 { 1 +1 } (2.1a)
2 - ,z
m1 + vp + m12' m2 + y['pc + m2C'
where p and m are the momentum and mass of the decaying particle. The symmetri-
cal decay angle 8 is also the minimum angle associated to a particular m and p.
A cut-off in the angle can therefore help to distinguish between different
particles.
A special case are decays with 3 photons in the final state, e.g.
J/~ ~ yX ~ y(X ~y). Combining (2.1) and(2.2) one obtains for the invariant
mass of the X-state
3. If not all of the decay products of a resonance can be detected (either be-
cause they are neutral or do not fall into the acceptance of the spectro-
meter), it is still possible to determine its mass if the resonance in question
R2 (fi.1.7c)is produced together with one other resonance Rl or particle.
From the masses mi and momenta Pi of the decay products of Rl the recoiling
mass can be calculated.
220 H. SCHOPPER
(2.4)
The results of such experiments are summarized in Table I. The way how masses
and the quantum numbers were determined will be discussed below when the produc-
tion and decay mechanisms of the new particles is described.
X
1 3p 3414 4
Pc 1 3p 3508 4
1
1 3p
____ 2_ 3552 6
- - - - - - - - - - - -- - - - - - - -
X
------ ---------
lji' , 3 3S ? 4028 1 30 3.772 3
1 1
- 4150
4414 5
0 IS 1863.3 0.9
0
0+ IS 1868.3 0.9 MARK 1104 )
0* 3S 2006 1.5
1
OH 3s 2008.6
1 1.0
F+ IS 2030 60
OASp105)
FH 3S 2140 60
1
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 221
We now want to compare the experimental mass spectrum of the cc- system
with the theoretical expectations based on the simple models described in
chapter 1.2. For the Hamiltonian (1.2) one expects a hydrogen (or positronium)
-like level scheme, as shown in fig. 2.1. One has different ladders for diffe-
rent angular momenta ~ and the levels are split due to hyperfine splitting,
spin orbit and tensor couplings. For a pure Coulomb potential the 2s and 1p
states would be degenerate. Because of the confining term (1.3) in (1.2) the p
and d-states are shifted to lower energies. Indeed for a r 2_potential (harmonic
oscillator) the 1p-state would lie in the middle between the 1s and 2s-states.
The lowest d-state has JF = 1- and hence could interfere with the 2s 1-state,
if the shift is large enough.
Ortho tt Para tt
35 J=r (1jJ') - - - ----
~ )=2+
J--O- 2p
$ =1
~ _
---'----<..s-o-\.:J:. $=1 1=3
J:1+ ~
s:-o-YI
J=2" -
1
~
$=J=r(~) t HFS
$ =J =0-
l= 0 2
Fig. 2.1: Level scheme for charmonium
222 H. SCHOPPER
Soon after the discovery of the J/~ and ~' particles several authors
solved the non-relativistic Schrodinger equation with the potential
as r 2
V(r) = - -
r
{1 - (-)
ro
} (2.5)
Identifying J/~ with the 13S1 and ~' with 23S1 state the following parameters
were found 5)
These values justify qualitatively the assumptions on which the naive model
was based: the Coulomb-like part is indeed short range, the coupling constant
is smaller than 1 and the c-quark mass is large (non-relativistic Schrodinger
equation). On the basis of these parameters the masses of higher lying non-bound
resonances were predicted 14 ):
Keeping in mind the simplicity of the model, these predictions are extraordinarily
impressive. The difference between prediction and experiment for the two highest
states can be attributed to neglecting coupled channels 14 ) (see 1.232).
Because of the spin-spin coupling the states with opposite and parallel spins
are split. If the X-particle found by OASP is identified with the 11so state (usu-
ally called nc) and the resonance at 3.45 GeV with the 21so state (n c ') (table 1a)
one finds the splittings
These large splittings have been considered for quite some time as a major diffi-
culty for the charm model, since most calculations produced much smaller
spl ittings.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 223
If the spin-spin coupling is associated only with the short range part of the
potential, one finds
where m1,m2 are the masses of the bound quarks and 1jJ(0) is the wave
function at the origin.
b) for general short range potential V(r)7,8)
short _ 2 2
6MHFS - 3 m1m2- < V V( r) > (2.7)
c) If spin-spin coupling is assumed to exist also for the long range part
of the potential, one obtains 18 ) for V(r) = ar
6M lon g = i (1 +de1) (1 +2) a <l> (2.8)
HFS 3 m1m2 r
Here anomalous couplings~, ~2 (see 1.232) for the long range force have been
included. The total splitting is
g~g!_~Eli~~i~g_2f_sb2r~2~i~~_~~2~~~
The experimental ratio RHFS = 6(1jJ-n c ) /6(1jJ' - nc ') = 1.2. According to (2.6)
one has RHFS=I1jJ(O)I~/I1jJ(O)I~,. The functions at the origin of the J/1jJ and 1jJ'
particles can be inferred from their leptonic decay widths and one finds
RHFS =r1jJ(e+e- )/r1jJ,)e+e- )=3.9/2.4=1.6. The agreement is not so bad and supports
the simple model.
For the absolute values of AM HFS on the other hand one calculates much
too small numbers on the basis of the standard model. Splittings of about 30
to 80 MeV were obtained 2,3,4,7) depending on various assumptions and on diffe-
rent values of ~(O). Using a more general potential V(r) the splitting could be
increased 8 ) to A(~ - nc) = 120 MeV and A(~' - n~) = 92 MeV, still too small by a
factor of 2. With special potentials the right values could be derived but this
leads to unnatural conditions for the lept0nic decays19).
Schnitzer 18 ) assumed that this coupling is small for the light quarks (~,
~d ~ 0) but appreciable for the charm quark (at c : 1). With such a large ~
the HFS-splittings and also the LS-splitting of the P-levels come out some-
what too bi9 11 ) *). A value Zc '" 0.4, i.e. (1 + rc)2 2 in equ. (2.8), reproduces
the experimental results better.
terms give 12 ) l+Z"'l + (4/3) (Ct/211) = 1.04, which is much too low. The matter
gets quite complicated if the long range behaviour is included. In that case
divergencies appear and arbitrary cut-offs have to be made.
The splitting between the charged 0+ and the neutral 00 meson has been esti-
mated by various authors. If the isospin breaking is calculated with a non-rela-
tivistic mode1 23 ) one finds 0+_0 0 '" 15 MeV, whereas a more refined mode1 24 ) yields
a value of 6.5 MeV in excellent agreement with the experimental value 5.1 2.8
(see table 1b).
Let us now turn to the splitting of the triplet P states which is associated
to LS and tensor couplings (see 1.23). For these states the quark spins are parallel
and therefore the total spin S = 1 couples with the orbital angular momentum ~ = 1
to total spins J = 0, 1, 2. From equation (1.5) one can derive the following ex-
pressions for the masses of the triplet P-states:
226 H. SCHOPPER
3 2
P2 =A + B - "5" C
3 (2.12)
PI = A - B + 2 C
3p = A - 2B - 4 C
o
A arises from the SS-term and the spin independent terms. It is useful to
define the ratio
This presents a serious difficulty since this value is outside the theoretically
acceptable range. From the experimental splittings and (2.12) one finds C/B ~ 0.3
implying that tensor forces have to be taken into account.
Difficulties arise not only for the ratio Rp but also for the absolute values
of the splittings. Associating the spin effects only with the short range Coulomb
potential gives splittings which are too small by more than a factor of 53). Includ-
ing the long range potential one obtains 35 l4 ) about the right value for the 3P2- 3P l
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 227
difference but the ratio Rp comes out wrong. The coupling to decay channels 14 ) and
anomalous moments have little effect (see 1.232) and cannot explain the big dis-
crepancy in Rp'
The only way proposed so far to remedy this difficulty is the assumption that
the long range potential L(r) is not a Lorentz-vector but contains scalar contri-
butions 21 ) (see 1.232,equ. 1.7). In this case the ratio Rp can have any value. Indee
for a pure scalar L(r), i.e. Kv = 0 in equ. (1.7), the ordering of the 3p states is
reversed, with the 3Po being the highest state, unless unreasonable values of as
are permitted 22 ). The experimental ratio Rp ~ 0.5 is reproduceo21 ) with
Ksl (Ks + Kv) ~ 0.8 implying that the long range potential is mainly scalar. It
should be noted, however, that Ks F 0 reduces the HFS splittings.
The 13D1 state is expected to be suppressed by spin effects and coming close
to the 23S1 these two states are likely to interfere. This will be discussed further
in chapter 4.1.
b) The potential Vir) = as/r + ar describes quite well the position of the
bound charmonium states as well as the charmed particles. Indeed, some
spectacular predictions could be made. Corrections due to coupled decay
channels and virtual transitions are not negligible.
228 H. SCHOPPER
c) The HFS-splitting for the charmonium states and the D and F mesons can
be understood in a common way if anomalous gluon ("magnetic") moments
ilre introduced for the long range force. Alternatively this splittino
mi9ht be explained in terms of instantons.
d) The experimental splitting of the triplet P-states can be reproduced
by L.S and tensor couplings, however, the long range potential has to be
a mixture between Lorentz vector and scalar, the latter being predominant.
Diagrams for the various possibilities are shown in fig. 3.1a and b.
The ce states are bound by the exchange of many "soft" gluons. Since the energy
of the gluons is low their coupling to the quarks given by as(E) is large
(see 1.22).
If gluons emerge from the cc annihilation they couple to light quarks. Each
of these gluons carries an appreciable fraction of the total energy ("hard
gluons") and hence the coupling as(E) is small. As a consequence these processes
are suppressed and the hinderance factor is determined by the number of exchanged
gluons (again in complete analogy to QED).
230 H. SCHOPPER
Hinderance
Example. factor
2
c
J~r } -e+ e~ 11+11-
:
c
:
~:: Ta
~
c
! : J/lil - ZIt,41t .... la
3
C
q~
lra
~ _
c _
c
c
0.-
J/IIIII--I-l1 c(O-+I+Y 3
a)
!
c
c
!
< l1clO+ -1- YY
Example Hinderance
factor
~~\f
3
JlIII- 31t as
, '- S::.
c ,. __ ,.;/5
~ __ ' I
J/III-~ZIt a~
c ',~q
q
~' J/lil - Y+ 11
Y+ 11'
a a~
A'
C ;\'q
'0.
0 111'- J/lil +11
J/IIIZIt
a s3
Vector particles with J PC = 1-- couple to the photon and hence can be
produced directly in e+e- annihilation (see Fig. 1.7b). As is well known the
IN, besides having been detected in p + Be .... e+e- + anything at BNL 25 )) has been
found at SPEAR 26 ) and soon after the ~,27). The decay of these particles was
investigated extensively at SPEAR 28 ) and at DORIS 29 ). Since detailed summaries 28 ,29)
have been published only the main points will be considered here.
~~~!_gg~gr~2~~~2Q~_Qf_~Q~~1_~~2_1gp~Q~2~_2g~~~_~22~b~
The width of the resonances turns out to be much narrower than the experimen-
tal resolution determined by the momentum smearing of the colliding e+e- beams
whi ch is about 1 to 2 MeV. But the true wi dths can be determi ned by a "tri ck" .
Assuming that the production cross section can be described by a Breit-Wigner
formula we have for the process e+e- .... J/~ .... final state
(3.1 )
where m is the J/~ mass, J its spin, ree' r f are partial and r the total
decay widths.
Integrating (3.1) over the energy E one obtains (with J 1)
(3.2)
If the three final states e+e-, ~+~- and hadrons are measured independently,
one can solve for the widths ree' r~~ and r = r h + ree + r~~. Since ree and
r~~ are found to be small compared to r h one has r h : r and from (3.2)
r (3.3)
ee
lji/J lji'
Lh (]Jb MeV) 10.4 1.5 9.7 1.2 9.6 1.7 3.7 0.6 3.06 0.34
L
ee (nb MeV) 790 965 141 790 200
L
]J]J (nb MeV) 870 100
r tot (keV) 69 15 87 20 67 25 228 56
*) corrected value
'I'll') assuming ree = r]J]J
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 233
The most direct evidence for the photon-like quantum numbers of J/~ and ~'
is an o~servation of the interference between e+e- 7 y 7 ~ +~- and
e+e- 7 ~7 /~- . The cross section is given by
d 911 (1 2~ ree I 2
dGo -_ BtL + cos 28- ) I - j + u
=-m------,.-E-----,.ir~77'i"2 (3.4)
- E~pected Interference
o Interference
(0) o(3095)
b
.
'- 0.1
::I..
b::l..
0.01
3.000 3.090 3.095 3.100 3.105
:: 0.10
b
'-
l
I::i 0.05
o L -__ ~ __ ~ ____L -_ _ ~ __ ~~
Fig . 3 . 2
234 H. SCHOPPER
e and ~ emission.
than 30 decay channels J/~ + hadrons have been observed. They will not
r~ore
The main conclusions drawn from the hadronic decays of the J/~ are the
following. The J/~ decays preferentially into final states with an odd number
of TI. The relation C = (-1) I . G together with G = C = -1 implies 28 ,29) iso-
spin IG = 0-, excluding I = 2 by the observed decay IN+pp. The fact that IN
also decays into an even number of pions violating isospin can be understood
quantitatively on the basis that the J/~ couples to the photon.
The isospin and G-parity of ~' can be inferred from the cascade decays
~' + IN + TITI and ~' + IN + n which account for 57 8 % of the ~' decays.
For example the experimental ratio (J/~oTIo) / (IN+TI+TI-) = 0.49 0.09 has to
be compared with the theoretical predictions 0.5, 0 and 2 for the isospin of the
TITI system I = 0,1 and 2, respectively. So clearly the pions have I = 0 and con-
sequently ~' and J/~ have the same isospin. The close similarity of the J/~ and
~' particles is obviously demonstrated by the fact that ~' + J/~ + TITI make up
about half of all the decays, whereas ~' + WTITI which has much more phase space is
about two orders of magnitude rarer. Also ~' + J/~ + TI o being I-forbidden is not
seen whereas ~'+ J/~ + n is allowed and was observed.
Finally it can be shown that J/~ and ~' behave as singlets with respect
to the approximate SU(3) symmetry of the 3 light quarks which is expected for
a charmonium state. The decays IN+ K+K- or KORo are forbidden for a SU(3)
singlet but allowed for an octet state. Indeed the experimentally observed
branching ratios for both the J/~ and ~' are very small (~ 10- 4 to 10- 3). The
SU(3) singlet nature can also be inferred from a comparison of J/W+ TIp and
KK*(892). In particular DASP results 34 ,29)indicate that the octet admixture
is very small.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 235
Finally one might ask if we understand globally the decays of J/~ and ~'
or if major decay components are still unknown. By summing up all known decays
of J/~ and adding those channels which can be estimated by I-conservation one
arrives at about 70 % of the total decay width. It does not appear unreasonable
that the major part of the missing 30 % is due to decays involving n's i.e.
J/~ ~ n + anything about which very little experimental information is available.
With respect to the ~' decays a little more than 50 % are associated with
cascade decays ~' ~ ~ + hadrons. The decays to intermediate P-states (see 3.3)
sum up to about 30 %. Adding the figure for electromagnetic decays (- 5 %) and
direct decays into hadrons (- 10 %) one arrives 28 ) at a total of about 95 12 %.
Not much room is left here for unknown channels, but again some rare undetected
decays might be of interest.
~;g~_Qi~~~~~iQ~_Qf_l~~!Q~i~_9~~~~~
A 1 -particle can convert to one virtual or 3 real photons. One real photon
is forbidden by momentum conservation and 2 photons by charge conjugation. Since
the transition via a virtual photon is of lower order it dominates.
(3.5)
where Q and m are the charge and mass of the quark involved and ~(O) is the
wave function at the origin (neglecting corrections as from annihilation channels).
Taking into account 35 ) "gluonic radiative corrections" equ. (3.5) has to be mul-
tiplied by {1 - 16 as / 31T}. For a general potential (see 1.23) I~(O) 12 has to be
replaced by m< dV/dr >.
In table 3.2 the experimental results of ree are shown together with ~(O)12
236 H. SCHOPPER
+ -
Table 3.2: Vector meson decays into e e
ljJ, ljJ' and ljJ'" shows. A more detailed analysis for the excited states has been
carried out 6 ) for a harmonic potential and including Sand D-wave mixing.
The regularities found for the leptonic widths of the vector meson ground
states are useful in identifying new vector particles as the I (see chap. 6).
~~~~_Q~!:fQ~~igg~~_b~g~Q~if_9~f~~~
u
allowed forbidden
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 237
Here the decay is shown as an example. The decay -+K+K- is allowed but
-+n+non- is associated to a disconnected diagram and hence suppressed. This
explains why the K+K- decay dominates in spite of its smaller phase space.
~~~~!_~!~9!~_9!~~Q~~~~!~9_9!~9r~~~
A J = 1 state can decay to 3 real photons as in the case of ortho posi-
tronium. In analogy we expect that it can also decay into 3 gluons. (In the
electromagnetic case a transition to one virtual photon is also possible.
One virtual gluon is excluded, however, because of conservation of colour charge.)
Using the well-known expressions for positronium
electromagn. - 64.2 1
r( 1 -+ 3y) = 9 (,,-9) fii7 (3.6)
and replacing a by as' also applying a colour factor is (see chapt. 1.22) one
obtains 2)
strong (3.7)
This offers a possibility to determine as' Assuming that the conversion of the
3 gluons to ordinary hadrons goes practically with 100 % probability
f(l--+ 3g -+ hadrons) can be identified with the full hadronic width if connected
diagrams are forbidden by energy conservation. This is the case for J/~. In the
case of the only the decays into non-strange hadrons must be taken into account
and for the ~I the cascade and the radiative transitions have to be discarded
since they do not proceed via a 3 gluon intermediate state.
Table 3.3:
M (GeV) r h/ r ee as Remark
(jl 1.1 588 0.47 (jl -+- non-strange *J
hadrons
J/IjJ 3.1 14.4 0.22
*) The reason for the small total width of ~ is that the OZI-allowed decay
~ -+ KK is suppressed by the small available phase space.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 239
Table 3.4:
From this table one infers that indeed ~' ... J/~1[1[ involving only 2 gluons is less
suppressed than J/~ ... hadrons**). One cannot expect more than a very qualitative
agreement since the gluons in the two cases carry different energies and hence
as may differ. One further sees that <P ... P1[ is less suppressed thanJN ... hadrons
but the reason is quite different. Here it is the lower energy resulting in a
larger as that causes less inhibition.
~~~1~ __ g2~~l~_9!~~2~~~~!~9_9!~gr~~~
The decay of J/y permits even more detailed studies of the mechanism of the
OZI rule. Besides the singly disconnected diagrams discussed in the preceding
section there exist also doubly disconnected ones. In fig. 3.3 different diagrams
are shown which give rise to the decays J/y + w~~, wKK and J/y + ~~, KK.
The left and right columns contain the singly and doubly connected diagrams,
respectively. The two diagrams in the first line show the w~~ and ~~
decay without intermediate resonances, whereas the diagrams of the other two
lines involve the tensor (2++) particles f and f'. The f contains predominantly
u and d quarks and can therefore be connected to the w whereas the f' is made up
essentially of ss and hence prefers to decay to KK.
An interesting question is now whether the fourth gluon in the doubly con-
nected diagrams (not attached to J/ y) is also a "hard" one implying a suppression
a~ or whether it is comparatively soft yielding a hinderance - a s3 An analysis of
the experimental data with this aim is complicated, however, since the influence
of the intermediate resonances has to be determined. For this reason the overall
ratio 28 )
>
J/-==f:'~
'-' f_
r:;-K
K
These results imply that most of the wn+n- decay width is due to resonance de-
cays. If these are subtracted the branching ratio for non-resonant wnn is of
the order of 0.1 % and hence comparable to nn. This is confirmed by selecting
only wnn events with M(nn) > 1.5 GeV, thus reducing resonance contributions, and
one finds B R(J/~nn) < 0.12 %. These data seem to indicate that the doubly
disconnected nn diagram is not much more suppressed than the simple disconnected
wnn.
f PLUTO
50
o~~~~~~~~~
o 1.~ 1.0 H 30
Invariant MassM IGeVlcl) .
10
o0~5'-'-'-L.f1.;;'-"-'
0~15"""""""'1""'
.0 ...........U,H,........uJ
3.0
Invariant Mass M~eV/cl]
e+e -.. JI $ +8 +n- .. (Oln +),,- .. (n+n - no) ( n+n-)
.. ",f .. (,, +,, -,,0)( ,,+,, -)
Fig . 3.4
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 243
~~g?_~~9!~E!Y~_9~~~~~_Qf_~L~_~~9_~~_EQ_Qr2i~~r~_b~2rQ~~
In this section we shall discuss the radiative decays J/w ~ny, n'y, nOy
and fy. From these reactions interesting information on the admixture of cc-states
to the ordinary hadrons can be inferred. On the other hand radiative transitions
from J/~ and w' to other charmonium states (pseudoscalar and P-states) will be
dealt with in chap. 3.3 and 3.4. Here the interpretation will be quite different
since these transitions take place between practically pure cc states. In order
to be able to discuss the decays like ny, n'y a few ideas concerning the mixing
of quark states have to be recalled (3.252). The puzzle to be explained is the
experimental fact that the decays ny and n'y are more than an order of magnitude
more probable than JN~nOy.
~~g?!_~~E~r!~~~E~!_2~~~~_r~E~~_(~rl_~~r_E~~~!~2_
The decays J/w ~n~ n'y, nOy and fy are two-body decays and hence photons should
be monoenergetic. The search for narrow lines in the inclusive photon spectrum
at SPEAR was negative 39 ), but the first 3 decays could be detected at DORIS by
investigating the decays with 3 photons in the final state40 ,41), e.g.
J/w ~ ny ~ (yy)y but also in the decay42) J/w ~ yn' ~ y (ypo) ~ yyn+n-. The
corresponding invariant mass plots M(yy) show the nand n' peaks (fig. 3.5).
and M(n+n-y) sho\'1 (fig. 3.6) a clear n' peak provided f1(n+n-) is restricted
to the p region.
30 11
t
~20
::e:
-
c:.
'"on
C
.....~
10
11'
~o .o
~ 30.0
is
~ 20.0
...~ 10.0
0.00 0
O.S 1.0
These figures are quite surprisi.ng at first sight. In terms of vector do-
minance one would expect that r(J/1jJ ->-YTIo) '" (a/y p 2 ) r(J/1jJ->- pOTI o )::: 1 eV with
r(J/w ->- pO,,0):::260 eV. This order of magnitude agrees with the experimental finding,
but why are r(J/w ->- n y) and r(J/w->-n'y) almost two orders of magnitude larger with
r(J/w->-n'y) ",r(J/w->- pOTI o)? And why is the ratio r(n'y) I r(ny) = 3.5 1.0 so
large?
A similar argument applies to the J/w ->- fOy decay which by vector dominance
is related to J/w->-fow with a measured 37 ,38) branching ratio of (0.40 0.14)x 10- 2
Hence the rate for the final state fOy should be comparable to TIOy but experi-
mentally it is much larger.
As will be discussed in the next section these anomalies can be understood in
terms of cc admixture to ordinary qq states.
Finally it should be mentioned that the corresponding decays w' ->- ny, etc.
have been looked for but not found 42 ,27).
~:~~~_I~~_~!~!~9_Qf_g9_~~~~~~
Thenyand y'n puzzle and some other experimental results (see chap. 3.3)
can only be explained in terms of mixing between cc states and qq states of
ordinary quarks. This mixing can be understood by extending the description of
the mixing of ordinary quarks.
It is well known that states of the ordinary vector and tensor mesons are
well segregated according to quark flavour. The neutral mass eigenstates exhi-
bit "ideal mixing" with (1/12~ ( uu ad) and 5S. As a consequence one expects
M(w) = t4(p) and M() = 2M(K*) - M(p). Both relations are fulfilled on the per-
cent 1evel .
The pseudoscalar mesons on the other hand show a strong mixing between
non-strange and strange quarks in spite of their different masses. As a conse-
quence the relations M(n) = M (lQ and ~1(n') = 2M(K) - M(TI) are strongly violated,
548 MeV against 138 MeV and 958 MeV against 854 MeV, respectively. The proper
masses are obtained with wave functions 44 )
::: -1 (uu - { + } .2
- + dd) ss
r,;!- (3.10)
/2'
246 H. SCHOPPER
which amusingly implies that the probability of finding a 5S or a uu/ad pair in the
nor n' meson is approximately equal.
The different mixing for the vector and pseudoscalar particles can be under-
stood 3 ,45,49) qualitatively in terms of gluon exchanges (see 1.22). To the mass
'matrix_correct~on termsAij have to be added Wh~Ch t~ke into account annihilation
terms qiqi ~ qjqj' If the 2 quarks are in a J = 1 state (vector particle) the
annihilation can only proceed via '3 gluons because of conservation laws. For a
JP = 0- state (pseudoscalar particle) annihilation via 2 gluons is possible. In
terms of QeD perturbation theory the first process is proportional to a~, the
latter to a~. For sufficiently large masses as < 1 and as a consequence the anni-
hilation corrections will be much more important for pseudoscalar particles than
for vector states, for which the Aij are negligible. Thus the different mixing
for the two kinds of particles can be explained in terms of QeD, but of course
one should not expect too precise results at low masses where as is still big.
If the charm quarks are included in the discussion the niass matrix for
pseudoscalar mesons can be written in lowest order perturbation theory:
Isospin symmetry gives AUU = Add = Aud' Since Mc ~ 3 GeV one may assume
AI M2c 1 and in a first step to determine the A.. the annihilation q.q.+ cc
lJ 1 1+
may be neglected. With the experimental masses of 'IT, K,n and n' as input one can
solve equ. (3.11) and finds 46 )
(3.13)
The immediate consequence of equ. (1.1) should be "uu > "us > \s' as indeed is
found empirically. Furthermore, the elements "ij should satisfy the factorization
relation "iiO"jj = "ij 2. From (3.12) one finds ("uu "ss)1/2 = 0.19 in good agree-
ment with "us = 0.21.
As a next step one tries to predict the annihilation corrections for charm
quarks 46 ). From table 3.3 one takes a s (Ms2) / as (Mc 2) 2 which is also in
approximate agreement with equ. (1.1). Thus one obtains
(3.14)
Hence as expected the annihilation corrections for the charm quark are
smaller than for the ordinary ones and consequently M(n c ) ~ Mc to a very good
approximation.
(3.15)
n = cc + e: n + e: I n I (3.17)
c
with
Estimating the coefficients Ev and Ev ' from the decay of J/~ into ordi-
nary hadrons one finds E , E' ~ 10- 4 . As mentioned above such a small admixture
v v
for the vector charmonium states is exactly what one expects since in this case
the annihilation corrections are associated to 3 gluon exchange and hence
A~.
lJ
- ra
[s
(m.) . a (m .)] 3/2
ql s qJ (3.19)
implying that the A~j are about one order of magnitude smaller than the Aij for
the pseudoscalar states.
~~g~~_Ql~~~~~lQ~_Qf_r2~12~lY~_~~~2~~
The decay (1 -- ) ->- (0 -+ ) +y is associated with an electromagnetic Ml transition.
If the quark content of the initial and final state is the same the transition
probability is given by (see also fig. 3.8a):
where Q and mq are the quark charge and mass and k is the photon energy. The
overlap integral n is expected to be of order 1 if the transition takes place by
a spin flip without changing the other quantum numbers. Indeed for transitions
with ordinary quarks like w+rr0y, p+TIy, +ny, KO*->-KOy one finds values of
n between 0.6 and 0.9. Transitions with charm quarks in the initial and final
state like J/1J!->-nc + y will be discu~sed in chapter 3.3.
Let us now discuss the case with charm quarks in the initial and ordinary
quarks in the final states like J/~ ->- ny, n'y, nOy and fOy. Since the emission
of the photon changes only the spin of the system but not the quark content such
transitions cannot go via diagram a) of fig.3.8. A mixing of states has to
take place which can either happen after (Fig. 3.8b) or before (Fig. 3.8c)
photon emission. As a consequence the overlap integral n in (3.20) has to
be replaced by E n where E is one of the mixing parameters of equ. (3.17)
or (3.18).
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 249
a) q 1-
q
/Y O
- q
q
b) c
C iJj(n
4{~~
11 c (o-J 11 ,11 I (0")
Fig. 3.8
~~g~~!_I~~_~Yl_~~Y_~~Q_fY_fl~~!_~~~~~~
In the previous section arguments have been given that the admixture of
cp and w to J/1jJ is very small (, ' ::: 10- 4 ) whereas the admixture of n, n'
to nc is of the order of 10- 2. Hence diagram 46 ,49) c) in fig. 3.8 can be neg-
1ected and transitions I ike J/1jJ .... ny, n'y will go by diagram b). In this case
one expects from equ. (3.20):
(3.21)
with the , ' values given after (3.17). The experimental ratio (see Table 3.5)
is 2.9 0.8 in good agreement with (3.21). Thus the unexpected ratio between
these two decays becomes plausible.
(3.22)
2 ~ 0.01 a width r(J/~ ~ ny) ~ 54 eV which agrees very well with the measured
width (see Table 3.5) of 55 12 eV. For JN~ n'y the corresponding values are
220 eV and 160 50 eV. These excellent agreements might be fort~itous in view
of the theoretical uncertainties but they indicate that lowest order QCD permits
one to understand the experimental results qualitatively.
The situation seems to be very similar for the JN~ fOy decay. Its bran-
ching ratio is of the same order as the JN~ fOw and hence vector dominance with
w ~ y as the source of the photons must be discarded, since it predicts a
fy/fw ratio of order ~y/y2w ~ 10 -3 . Again a diagram of the type b) in fig. 3.8,
where the photon is emitted by the cc must be dominant, this time leading to a
cc state with J PC = 2++ which then mixes with the f. A quantitative analysis 51 )
has recently been carried out. There it is shown th&t the angular distribution
of the photon contains interesting information.
~~g~~g __ I~~_9~~~_~L~:_~~y
For this decay the situation is reversed. Since nc - ~o mixing is forbidden
by isospin symmetry it is diagram c) of fig. 3.8 which is dominant and b) is
neglegible. Since the J/~ -w mixing is quite small, as explained in 3.2552, it
is plausible that the rate JN~~Oy is much smaller than JN~ny, n'y.
= 41 (3.24)
In conclusion one might say that first order QeD calculations provide a
consistent picture for the radiative decays of charmonium into ordinary hadrons
explaining in particular the drastic difference between J/~ ~ ny, n'y and nOy.
A study of these decays gives simultanously a deeper insight into the mixing
of qq states.
Of course besides QeD there are other possibilities to understand the data.
Schemes involving SU(4) breaking have for example been developed 16 ,48) which do
not need an explicit potential. Many relations between various decay channels
could be deduced.
~~g~~~ __ Q~!_~~~~~1~12~~_~1~~2~~_~~_:_99_~~9_~~_~j_~_Q
If the transition rates are calculated49 ) taking into account only SU(3)
singlet and octet mixing for nand n' one obtains for all forbidden diagrams
with 3 gluon exchange (fig. 3.8c) for the final states the ratios
with the mixing angle 8 ~ 110. This is obviously in drastic disagreement with
experiment.
The 2 gluon diagram (fig. 3.8b) yields
Here the enhancement of ny and n'y with respect to nOy is properly reproduced
but the ratio r(n'y) / r(ny) cot 2 e ~ 30 is much too large compared to the
experimental value 3.1 1.
It seems that the decay ~' ~ J/~ + n is more analogous to the radiative
transitions than usual all forbidden decays. Its width 27) r (~' ~~n) = 9.3 1.6
keV is very large for its little phase space (Q = 40 MeV) and being p-wave, SU(3)
and all suppressed. Hence this decay seems to confirm the cc ++ qq mixing as
established in the radiative decays.
252 H. SCHOPPER
In terms of the charmonium picture one expects that the vector particles
J/~, ~' etc. with parallel quark spins (3 S1 states) are accompanied by pseudo-
scalar states with opposite quark spins (IS states) (see chap. 2.2). These
pc _+ 0
states have the quantum numbers J = 0 . Since they are analogous to the
n particle of the light quarks these states are usually denoted by nc (= 11So)
and nc ' (= 21 So ). Because of the hyperfine splitting one expects these states
to be somewhat below the J/~ and ~', respectively.
3.31 ~~E~~!~~~!~!_~~~~!!~_!Q~_!b~_01~~~~2_~~9_xl~~~~2_~!~!~~
Since the pseudoscalar states have even C they cannot be produced directly
in e+e- annihilation but can only be reached by decays from J/~ and ~'. Since
these decays are rather weak our experimental information on the nc and n~
states is very scanty compared to the rather complete knowledge of the vector
states J/~ and ~'. Indeed the existence of the nc and n~ has been in doubt for
quite some time. There is positive evidence now for nc which is identified with
the X(2.83) particle detected by DASP, whereas the existence of the x(3.45)
level still needs definite confirmation.
OASP
30
-VI
C
<1.1
>
W
I
'0
/ I
I
o 10 / I
z /
I
I
I
I
/
./
."
" ....
Fig. 3.10
25.4 H. SCHOPPER
The branching ratio of the decay chain is given in Table 3.6. The DESY-Heidel-
berg 53 ) experiment sees also an excess of events in this mass region but be-
cause of their worse mass resolution they cannot give but an upper limit of the
branching ratio. The X particle has recently be seen in a pion-proton scattering
experiment at Serpukhov 56 ) and hence its existence seems to be well established.
The monoenergetic y-l i ne a ri sing from ~ I + n~ has not been seen (see
Table 3.6) as well as hadronic decays. The situation is similar to that for
the nco
~~~g __ Ib~Qr~~l~~!_~~~~~~~~lQ~_Qf_~r~~~l~lQ~_r~~~~
3.321 M1
- - transitions
- --- --
The decays J/~ncY and ~I+n~ are allowed magnetic dipole transitions.
The transition rate is given by
(3.27)
where Q and M are the c quark charge and mass, k is the photon energy and
n is the overlap integral of the initial and final wave function. Allowed M1
transitions are those between states which have essentially the same spatial
wave function and differ only in the spin state. Hence n~ 1 if spin-orbit
coupling and other spin-dependent effects are neglected.
With these assumptions one calculates 4 ,59) from (3.27) a decay width of
r(J/~+ync)~ 29 keV. From the branching ratio in table (3.6) the limit
256 H. SCHOPPER
Similarly one calculates 59 } the decay width r(~' -+ y n'} 17 keY whereas
c
the observed limit is r < 6 keY and hence Q2 < 0.3.
The two "cross" transitions n~ -+y J/~ and ~' -+ y nc are forbidden Ml
since the main quantum number n changes from 2 to 1. Hence the wave functions
are orthogonal if spin-dependent forces are neglected. The width corresponding
to the second term in the power series of exp(ikr} is given by
where Qforb =< f I r21 i> and the meani ng of the other symbols is the same as in
(3.27) .
Using different wave func':ions r(~' -+y nc} between 1 keY and 10 keY was
found 4 ,5,14} which should be compared to the experimental partial width
r(~' -+y nc} ~ 2.5 keY. This agreement, however, has not much relevance. It
The situation for the decay n~ -+y IN is more compl icated. From the data
given in Table 3.6 one calculates a branching ratio B(n~ -+ y J/~} >(O.30.16)
which is very large. Since the total width of n~ is not known it is not possible
to give an experimental partial width corresponding to this branching ratio. On
the other hand the theoretical estimates deliver only r(n~ -+ yJ/~} and not the
branching ratio. Because of the unreliability of equ. (3.28) a better "theore-
tical" estimate can be obtained by assuming that the matrix elements of n~ -+y IN
and ~'+y nc are the same and correcting for phase space one obtains
r(n~ +y IN} :: (1/4}r W -+yn c ) 1/4) . 2,5 keY '" 0.6 keY.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 257
3.322 nc ~ yy
- - - --
Transcribing the QED results for the singlet state of positronium one
finds for the decay rate 2 ,4)
(3.29)
Taking lj!(0) as determined for the J/lj! (chap. 3.23) one obtains r(n c ~ yy) '" 8 keV.
The experimental branching ratio B(nc~YY) > (8 3.5) x 10- 3 cannot be com-
pared to this expectation since the experimental total width of nc is not known.
Theoretical estimates will be discussed in the next section.
3.323 ~~9rQ~!_9~~~~_Qf_Dc_~~9_D~
Followin9 the analogy between QED and QeD the pseudoscalar nc and n~
can annihilate into 2 gluons besides into 2 photons. To derive the 2 gluon
annihilation rate from (3.29) one has to make the replacement 35 )
a 2 Q4 ~ 2a 2 s /9 and obtains
Inserting as = 0.2 and using lj!(0) as determined from the J/lj! one has
r(n c ~ hadrons) '" 6.4 MeV. This can be checked by an estimate which is independent
of the wave function and is obtained by taking the ratio of (3.30) and (3.7)
r(n c ~ g g) 27 7T
'" 100 (3.31)
r( J/1jJ ~ ggg )
258 H. SCHOPPER
--.:.i
3.9 x
(l:l)
3.7 2 0 . 43
With this theoretical hadronic width one can now calculate the branching
ratios B(Tl c +yy) = r(Tlc+n) / r(Tl~ -> hadrons) '" 8 keV/7 MeV", 1.2 x 10 -3 . If
this is compared to the experimental 1imit B(Tl c -> yy) > (8 3.5) x 10- 3 one obtains
a factor> (7 3) of discrepancy. In view of the uncertainties this does not
seem too worrisome.
A more serious problem is the fact that so far no hadronic decays of either
the Tlc nor the Tl~ have been observed. One way out could be that these particles
decay to many different final states which are difficult to identify. With sta-
tistical model calculations some authors 61 ) have tried to estimate the proba-
bility of different final states. They find that indeed each channel contributes
only a few percent. e.g. Tlc ->- 2 TI + 2 TI - 2TI 0 7 to 15 %, Tlc ->- 2 TI + 2 TI - , 3TI+-
TI
2 to 5 % each, Tlc ->- pp + anything a few percent. The first channel which is do-
minant is hard to find. The limits given in table 3.6 may still be compatible
with these rates.
3.33 Q!~~~~~i9~_9f_Q~~~99~~~l~r_~!~!~~
The comparison between experimental results and theoretical expectations
concerning the Tlc = X(2.83) and Tl~ = x(3.45) particles can be summarized in the
following way:
a) Whereas the existence of the X(2.83) is well established, the x(3.45)
needs confirmation.
b) The large splitting between the 3S and IS states which originally has
caused much concern does not seem too much of a problem. It can be
understood quite coherently by attributing an anomalous gluon coupling
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 259
to the quarks (see chap. 2.2). Recently it has been shown 62 ) that the
existence of instantons may be responsible for the large splitting and
indeed it could be considered as direct evidence for the instantons.
c) Serious difficulties exist for the Ml transition rates. In particular
it seems difficult to explain the large discrepancies for the allowed
Ml transitions between HFS partners. The even larger discrepancies for
the forbi dden Ml "cross transitions" are perhaps 1ess worrying because
of the large theoretical uncertainties.
d) The real puzzle is the fact that no hadronic decays of nc and n~
have been seen, although partial width of several MeV are expected.
The hadronic transitions seem to be suppressed and one puzzling conse-
quence is the very large B(n~"'Y IN) > 32 %.
Several authors l6 ,48,64) have shown that the transitions J/~ ... nc Y
and nc-+-rY can be hindered by SU(4) symmetry breaking. Indeed by a proper choice
of the parameters r(nc"'YY) < 3.7 keV can be obtained. Maybe relativistic
models are needed to interpret the decays involving nc and n~.
~~~!_~~e~~!~~~!~l_~~~~l!~_fQ~_~~:~!~!~~
The first evidence for an intermediate state called Pc was found by
DA5p ) in the cascade ~' -+ y Pc -+ y (y IN). Later this and cascades to other
66
intermediate states,which were given the generic name X} were seen at 5PEAR 54 ,55)
by DE5Y-Heidelberg 68 ) and PLUT0 69 ). The results are shown in fig. 3.11.
Clustering in the (IN y) invariant mass can be seen at 3.42,3.45,3.50 and 3.55
GeV. All these states have even C since they are reached by an electromagnetic
transition from a 1 state. The state at 3.45 GeV has a different character
than the others since its decay into hadrons has not been observed. As dis-
cussed in chap. 3.3 it is tempting to identify this state with the 2150 level
and since the relevant experimental data have been given in chap. 3.31 we
shall not discuss it here.
The mass distribution of the y J/~ system in the decay ~' -+ yyJ/~ as ob-
tained in recent measurements 68 )80) is shown in fig. 3.12 a and b. The peaks
at 3.51 and 3.55 GeV are clearly seen, there is a small indication at 3.41 GeV,
but no significant structure is observed at 3.45 GeV.
For the three P-states the monoenergetic lines corresponding to the transitions
~' -+y3 p have been observed (fig. 3.13) and the branching ratios were determined.
These branching ratios together with those for the yy cascade are collected in
table 3.7.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 261
Ps IX -status
l.I6eV
lA t ~' -Xy
Ly~
" lIS6.v L II'II-
~.
I
U\ .8 llll6eV
}ll
9
4>< .I
8SP.. J!l.i itY
~
"0 .
",,~g I
lIS
PLUlij
~
g
0
,
0
x DASP
3.2 o SLiCILBL "
JI lIS 11 m
Mllyl (Gevl
Fig . 3. 11
DESY-Heidelberg
NoIS MeV
15 .0 r - - -- r - - - - r- - -r - - -- - ,
DASP ~' ~ yy J/ ~ Iji- yyJ/~
....
N
U
myy< .S GfN
:>-
OJ
708 ewnts
'"
~ 10.0
....
a
....'"c:
OJ
>
OJ
'::; S.O
o
z
~l (y IN)
300
200
100
)...
w
"):: 0
w
<I
0~
300
rf")
a:
w 200
a..
(f)
I-
z 100
:J
0
U
0
80 (e)
40
- 40
50 100 200 400 800 2000
Ey ( MeV)
Fig. 3.13
Single photon spectra from the decay of ~' as observed by MPPSSS054)
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 263
Information about the spin of a 3p state can be obtained from the angular
distribution of the first (low energy) y in the cascade with respect to the
beam axis. One expects a distribution proportional to 1 + a cos 2 e where e
is the angle between the first photon and the beam axis. With J = 0 for the
intermediate state one has a = 1. For J = 1 or 2 the prediction for a is not
unique since mixing of multi poles is possible. The experimental results for
a are also shown in Table 3.7.
No 3p -+ yy were observed 69 ).
264 H. SCHOPPER
a) We use B(1jJ'-ryX) = 0.075. The errors quoted for B(X->f) do not include the
overall scale uncertainty of 35 % due to the error ( 0.0026) in B(1jJ' -> YX).
a) We 'use B(1jJ'-> YX) = 0.071. The errors quoted for B(X -> f) do not include the
overall scale uncertainty of 27 % due to the error ( 0.019) in B(l)!' ->YX).
a) We use B(1jJ' ->YX) = 0.070. The errors quoted for B(X -> f) do not include the
overall scale uncertainty of 29 % due to the error ( 0.020) in B( l)!'-> yx).
266 H. SCHOPPER
it to the 21S level as discussed in chap. 3.3. However, since neither hadronic
decays nor angular distributions could be observed there is no direct experi-
mental information on the quantum numbers of this state except its even charge
conjugation.
~~~~_I~~Q~~~2~~!_~~e~~~~~2Q~_fQr_~!_~r~~~2~2Q~_r~~~~
The transitions from ~' to the 3p levels and from these to J/~ are electric
dipole transitions.
For ~'~ y 3pJ one finds 4 ) for the partial width
(3.33)
where Q is the quark charge, k the photon energy and J the spin of the final
state.
The experimental values for the branching ratio B (taken from table 3.7)
and the partial width r J (calculated with rtot(~I) = 228 keV) are shown
in table 3.9. The width r J for the 3 transitions are remarkably constant. This
is due to the circumstance that (2J + 1) k3 which according to (3.33) deter-
mines r J turns out to change only by a factor of 2 (see table 3.9) because
of the balancing between (2J + 1) and k3 The factor k3 alone changes by a
factor of 50. As a consequence the experimental constancy of r J supports very
strongly the assumed spin ordering since for different spins r J would change
by large factors. This statement is independent of detailed calculations of
the matrix element.
Original estimates 2 ,4) of the absolute decay rate were based on too
crude approximations and lack of precise knowledge of the transition energies.
The results of later calculations are given in table 3.9. They are in very
good agreement with the experimental r J . This confirms that the c-quark charge
is Q = 2/3 since for Q = -1/3 a discrepancy by a factor of 4 would emerge.
The transition probabilities for the decays 3PJ ~ yJ/~ are given by
(3.34)
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 267
Tab 1 e 3.9
3p
~ 7 Y J transitions
Tab 1 e 3.10
Compared to (3.33) (2J + 1) has been replaced by 3 and hence no balancing of the
k3 dependence occurs. The experimental results for the branching ratios are given
in table 3.10. Since the total widths of the P-states are not known r J cannot be
determined. On the other hand, from (3.34) only r J can be calculated. As a conse-
quence a direct comparison between experimental and theoretical results is not
possible. Jackson?3) has tried to estimate the total widths from sum rules and
QeD but the results are not accurate enough to enable a detailed comparison.
Quantitatively one sees by inspecting table 3.10 that B(3 P1 .... YIN ) is lar-
ger than the other two branching ratios and also its absolute value of (34 11) %
is remarkably big. In contrast the theoretical value r J for this transition lies
between the other two. This indicates that the hadronic decays of 3P1 (3.51) are
suppressed compared to the other 23P-states. This can be understood in terms of
QCD as will be explained in the follwing section.
In terms of the simple charmonium picture the hadronic widths of the P-states
are given approximately by the annihilation of cc into two gluons 35 ) for J = 0,2
and into 3 gluons(or 1 gluon and a light qq pair?4))for J = 1.
(' ~ ~q
c c c
or
c c c
9
q
J =0,2 J =1
(3.35)
where M is the c-quark mass, R' is the derivative of the radial part of the
wave function and the numerical constants are No = 256/3 and N2 = 1024/45.
The 2 gluon annihilation is obtained by the substitution a 2 Q4 .... a~(2/9) where
Q is the c-charge and one arrives at
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 269
1 R'(O) 12 (3.36 )
Using the experimental PJ~yJN branching ratios one gets estimates for the
radiation widths of 70 and 60 keY, respectively. A comparison with the theore-
tical rJ(PJ~ y J/~) listed in table. 3.10 shows discrepancies of factors of 4
and 2. This can be considered as a reasonable although certainly not good
success of QCD.
The decay probabi 1iti es PJ ~ yy as cal cul ated from (3.35) are completely
negligible to the hadronic widths (3.37) and therefore also to the transitions
PJ ~ yJN . Indeed the two photOll decays have not been observed 57 ) with
B(~'~ PJy ~ yy y) : 4'10- 4 .
For the J = 1 P-states the QCD prediction is much less certain. It has
been argued 74 ) that the annihilation into g + qq dominates the 3P1 (J=I++)
decay whereas the ggg final state is most important for the IPI (J=I+-) state
which has not been observed so far. Both of these transition rates involve
logarithmic divergences at zero binding energy and
(:(3
r '"
an
P (1T )2 ()
1R (0) 12 (3.39)
C C
reason 5) is that the decays 3pJ ~ J/~ +rrrrhave little phase space and are hin-
dered by the centrifugal barrier.
An enormous effort at SPEAR and DORIS has produced over the last few
years convincing evidence for the charmonium model and for QeD.
IjI (4411
m:M$f~
FF*; OILOii*elc
t2
Mass.
GeV
18
3.6
34
12 21t or2K,4ltelc
3.0
2.8
t t
3P
t
30 1
1.2.3
Fig. 3.15
272 H. SCHOPPER
A measurement of the total cross section e+e- + hadrons can give impor-
tant information on quarks and the strong interaction. As discussed in
chap. 1.13 this process can go via an intermediate resonance state with J = 1
provided the total e+e- energy E coincides with the rest mass of the reso-
nance.
Away from resonances the total hadronic cross section can be estimated
in the quark model on the basis of an asymptotically free theory with gluonic
corrections and one finds 76 )
(4.1)
where Q.1 are the quark charges, a s is the gluon-quark coupling constant whose
energy dependence is given by (1.1), and
is the e+e- + ].l+].l- total cross section. The sum has to be extended over those
quarks whose qq pairs can be produced at a given energy E. The factor 3 comes
from the 3 colour degrees of freedom (see chap. 1.21).
A heavy lepton with a sufficiently large mass can decay into hadrons and
hence can contribute to o(e+e- + hadrons). If the lepton is pointlike its
production cross section will be equal to 0].l].l (36 - 63 )/2 where the term mul-
tiplying 0].l].l gives the threshold behaviour for a spin 1/2 particle (6 = vic of
the 1epton). The contri but i on to R will be 6R l epton = B x (36 -6 3 ) I 2 where
B is the branching ratio of the lepton into hadrons. Since, however, in most
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 273
experiments measuring 0tot hadrons, electrons and muons are not identified decay
electrons and muons are taken for hadrons and consequently B = 1 and
6Rlepton ~ 1 well above threshold.
R
4
2
,11,/11"
0
0 2 4 ~ 6 7 8
Fig. 4.1 E c..... (G. V )
to the opening up of the charm threshold. The resonance peaks appearing around
4 GeV will be discussed In chap. 4.2.
The looser trigger enabled PLUTO to break down the total cross section
into charged mUltiplicities. as shown in fig. 4.4. It is remarkable that
10
2.0
1.0
h!!k/'t.J{T.IJ1r
I
0)0.2
a
3.5 1,,0 4,5 5.0
3,0
2,0 I
~Vt
iI\IlfII.1Ill I
1.0 ,/-
b) n.4
0
3.5 4,0 4,5 5.0
2D
1.5
..
....-,;~~r.I.
J __ _II
1,0
0.5 cJ n,,6
0
3,5 4,0 4,5 5,0
W [GeV!
Fi g. 4.4
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 275
these partial cross sections and in particular the 2-prong cross sections
show the same structures as 0tot' The dip at 4.3 GeV is deepest in 2-prongs,
and indeed at this energy R can be attributed almost solely to light quarks
and the heavy lepton.
Fig. 4.5
In conclusion it can be stated that the data are in quite good agreement
with the 4-quark model and the existence of the heavy lepton and they confirm
expl icitly the colour factor 3 in equ. (4.1).
Above the threshold for two D-mesons Ethr = 2MD = 3.726 GeV charmonium
resonances can decay into DO. Above this threshold one expects therefore pro-
cesses of the type e+e- -+ ljJ* -
-+ DO according to the quark diagram
>------------~y-;--~--C~;=:=:: ~D
5
4.31 I~~_=g{~~ZZl:~!~!~
A resonance very close to the charm threshold was found at (3.77 6) MeV
by the SLAC-LBL collaboration 84 ). Its widths was found to ber= (28 5) MeV.
This can be considered a beautiful confirmation of the arguments based on
charm and QCD since the ljJ',which lies only 88 MeV lower, has a much smaller width.
This demonstrates the difference between an allowed and a OZI hindered decay.
The mass, the total and leptonic widths have been predicted by the Cornell
theory group14,85) for the lowest 3D1-state with the help of the potential
model (see chap. 1.23) with astonishing accuracy. Therefore one might ask why
this resonance has not been detected before. The reason becomes obvious from
fig. 4.6a) which shows the uncorrected data. One notices that the resonance
falls on the large tail of the ljJ' and an appreciable amount of statistics had
to be accumulated.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 277
rad corr b)
"*
later one calculates a DO threshold
at 3.87 GeV which is well above the
I~
resonance. As a consequence the 3.77
!! GeV particle should decay almost ex-
~, clusively into DB. The branching ratios
determined from the individual 0 decays
171 3 S~ lSS are listed in table 4.1. This property
makes the 3.77 resonance a clean source
Fi g. 4.6 for 0 particles whicn provides the
possibility for a detailed study of
these particles (see chap. 5.2).
(4.3)
where a and b are mixing parameters and ree (3 01 ) is the leptonic width of
a pure 301 state. With relativistic corrections r ee (3 01 ) ~ 0.1 keV was found 75 )
278 H. SCHOPPER
and inserting the experimental ree for the 3.77 and 3.68 GeV states one
obtains a/{a+b) ~ 0.13. This implies an appreciable mixing. If the mixing is
expressed in terms of a mixing angle one finds 84 ) e = (23 3)0.
State Mass r r
MeV/c 2 keY keY Branching ratios Experiment
30 3772 3 2B 5 0.37 0.09 0000 44 22 % SLAC-LBL82 )
1
0+0- 44 33 %
3770 6 24 5 0.18 0.06 OELC086 )
4035 2 55 5 0.7 0.1 PLUTO B3 )
4040 10 52 10 0.75 0.15 OASp 123 )
4146 4 47 11 0.4 0.1 PLUT083 )
4156 20 78 20 0.77 0.23 OASp 123 )
4414 5 33 10 0.44 0.14 SLAC-LBL 82 )
4400 3 33 9 0.3 0.1 PLUT083 )
4417 10 66 15 0.49 0.13 OASp 123 )
The total e+e- hadron cross section shows further peaks at 4.03, 4.15 and
4.40 GeV. If they are associated with resonances one can determine the parame-
ters given in table 4.1. These resonances could be 3S1 levels or 3D states with
some S-state mixing. Unfortunately there is no direct experimental evidence
to support such assignments. Also the predictions from the potential r,;odels
are not very reliable because of decay channels and other corrections (see
chap. 1.2). Structures can also be produced by opening thresholds, interfer-
ence effects between resonances and form factors in e+e- ~ DO, DO*, D* 0* etc.
Consequently the identification of structures becomes less reliable the higher
the energy.
The pronounced peak at 4.03 GeV is due to the conjunction of the 33S1 state
and the opening of the D*O* threshold at 4.012 GeV. That this peak is never-
theless a resonance can be seen in the Argand diagram for DO scattering I4 ,87).
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 279
The peak at 4.15 GeV could be attributed in the standard model to the 2301
state but it also lies just above the FF threshold at 4.06 GeV. To clarify the
character of this peak it would be interesti ng to look for F production which
is difficult, however. Nevertheless some evidence for F production has been
found 123 ) (see also 4.43). The Argand diagram supports the interpretation as
a resonance 14 ).
The resonance at 4.40 GeV finally could be associated 85 ) to the 43S1 state.
This is supported by the splitting 4.40 - 4.03 GeV ~ 400 MeV which agrees with
the expectation for 43S1 - 33S1 but does not fit so well with a O- state. A sim-
ilar argument applies to the leptonic width. The ratio ree (4.4) / ree (J/1jJ) ~ 0.1
seems to be compatible with an S-state but too large for a O-state (see chap.
3.21)85).
Some broad structure seems to become apparent around 3.95 GeV (see fig. 4.7).
This bump may have to do with the opening of the oD* threshold at 3.872 GeV and
with a zero in the 3S decay amplitudeS7 ). Similarly the broad structure around
4.35 GeV may be associated to the FF* threshold at 4.17 GeV. To check this it would
be very interesting to look for F and F* production in this energy region.
$L.AC-LBL SPI7
O~L-------~~~9--------~.D~------~.J
WIGeV)
280 H. SCHOPPER
Threshold Pea k
DO'" 3.872
- 3.95 threshold + form
factor
4.012
4.03 335
1
~O, DO"', o"ii*
+ threshold
FF 4.06
4.15 2301 Ft
or threshold
4.17
- 4.35 threshold?
4.28
4.40 Fr'"
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 281
The 4.03 GeV resonance is above the thresholds for DO,- DO-* and 0*'=0"* decay
but below the thresholds for decays into F-particles. At SPEAR a large sample
of data has been accumulated and the branching ratios for the 3 decay channels
(see chap. 5) have been measured 88 ,89). The results are shown in table 4.3. The
surprising result is that after correcting for the very different phase spaces
the 0* production is very much favoured. In a naive quark model the ratios of
the production probabilities should simply be given by statistical spin factors
which indeed favour the 0*5* channel but much less than experimentally found.
DO from DO 50
D 5*0 +D*5 D* 5*0 Charged 0 Ref.
Fractions 0.05 0.03 0.38 0.08 0.40 0.10 0.16 0.11 88) 89)
experiment
Exp. after
phase space 0.2 0.1 4.0 0.8 128 40 70)
corrections
Statistical
spi n fac- 1 4 7 85) 23)
tors
Node in decay 0.04 4 30 85) 87)
amplitude
Oecoupling 0.9 4 175 90)
scheme
The Cornell theory group has tried to explain the experimental ratios
in terms of a coupled channel model which produces a P-wave decay amplitude for
33S -+ (cu) + (cu) which is oscillating and has a node. As a consequence the cross
section a(e +e - -+ ~O)
-
has a zero near 4 GeV and the DO- production is strongly
suppressed with respect to O*D*.The ratios inferred from this model are
given (after phase space corrections) in the fourth line of table 4.3. The
282 H. SCHOPPER
tendency agrees with the experimental values, but there are still differences
of factors 4 to 5. While it seems possible to tune the model to get better
agreement on the ratios it seems difficult 88 ) to get reasonable ratios and at
the same time explain the maximum of R.
An alternative explanation 91 ) has been offered for the large D*O* pro-
duction by assuming that the 4.03 GeV particle is a D*5* 'molecule'. If this
were true, one expects by a rearranging of the 4 quarks the decay into J/~
which was estimated to occur on the 10 % level. PLUT0 92 ) has looked for the
inclusive production of J/~ in the energy region from 4.0 to 5.0 GeV. A
cross section of 31 21 pb has been found which corresponds to 0.13 % of
the total hadronic cross section. Consequently no enhancement due to 'mole-
cules' has been seen and the observed J/~ production can be understood 93 )
in terms of normal OZI forbidden transitions.
A step in K production at an energy ~Ihere the e+e- ... DD- channel opens
up provides very good evidence for the existence of charmed particles.
Strangely enough such an increase in the inclusive cross section could not
be seen for quite some time. Only at the beginning of 1977 was PLUT095} able
to see a step in the K~ cross section at an energy of 4 GeV. This was imme-
diately confirmed by DASp 96 } for charged K mesons (fig. 4.8). Assuming that
the increase found in the K cross
section has the same origin as the
step in 0tot one finds that 60 to 80 %
of the final states produced in the new
nb
.=ale+~ _K~ ...anythingl , PLUTO phenomenon contain kaons. The number
0: 1/2a(e"'e--K"! ...anything), DASP
of K~ per event is 0.39 0.06 and
since the number of all kaons is four
times larger one has 1.56 0.24 kaons
per event, which comes close to 2.
This figure is expected if pairs of
charm particles are produced and each
lO L5 5] decays into a kaon. Additional proof
Inclusive Cross Section of Kaons
that the kaons are associated to charm
Fig. 4.8 production comes from the K spectra.
284 H. SCHOPPER
PlUTO Kt
OSUC-l8l ..;
K'
Eo. [GoY]
Fig. 4.10
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 285
Three groups have measured inclusive electron spectra. DASp Bo )126) and
B6
DELCO ) use Cerenkov counters for electron identification, whereas I1ARK 1 98)
has recently be complemented by a lead glass wall. The results of the three
groups are presented in fig. 4.11. Indeed, one notices the rise of
Re = a(e +e- -+ e + anything) / a(e +e- -+j.J +-
j.J )
DELCO
R from zero below charm threshold to values
.~
1.1 II ~tj
ration of the lepton decays was not per-
fect and hence the value of Re at the dip
could be entirely associated to T decays.
# ----- ---------
_ .,i....
..
06''- .... ".
One can now attempt to calculate the branching ratio Be of the charmed
mesons into electrons:
r R'
e (4.4)
r 2 Rcharm
The results are shown in fig. 4.12. The SPEAR data tend to be lower than
the DASP results which is mainly due to the difference in R used to calcu-
1ate Be'
Fig. 4.12 demonstrates that Be does not change much with energy. This is re-
markable since at higher energies F mesons and charmed baryons can contribute.
The constancy of Be can either mean that F and baryon production is negligible
or that the branching ratios of these particles are similar to those of the
D meson.
~ 0.30,..,---,--,--.---.,---.-.--.--.-01""
-E
b 0.25
u
~ 0.20
l-
n.
~g 0.15
-oe{
~~
f/)(!) 0.10
w~
(!):I:
~!:!! 0.05
Woe{
~!f5 0.0'--'----.J_----1_--L._~_...i..._...L.._..L-_-41-..IJ
3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2 7.0
Ec.m.(GeV)
Fig. 4.12
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 287
Kaons in the final state indicate the decays of D mesons, the appearance
of n is a sign of F decays (see chap. 5). The inclusive yield of n offers
therefore an interesting possibility to find energies ~Ihere F production is
appreciable.
30
~
-
~
~
Kl
c
c
w 20
10
o 0.5 to
Myy in GeV
Fig. 4.13a
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 289
1,0
1,.00 O!E W O!E 4.10 GeV
>QJ
~
Lf)
10
.......
....~
c
w
10
0.50
Fig. 4.13b
290 H. SCHOPPER
0.4 ,---,--,----,--,----,---,-----.
-g 03
1
;>-0.2
F
f
~0.1
-----l------r-----------+
I:>
Fig. 4.14
30
>QJ
:L
oLf)
JJ
.~ 20
C
UJ
10
Myy in GeV
Fig. 4.15
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 291
PLUTO
RI,I 01..- --so.anll
a,.,. prelinVnary
t t tt
l
w
3.S 4.0 4.S 50 GoV
4.16
errors which are independent of energy. The data, therefore, exhibit a step
just below 4 GeV, the threshold of charm, but also of heavy lepton production.
The average value of R(po) = 1.3 for W> 4 GeV may be used to estimate which
fraction of charged pions comes from vector mesons. The result is that more
than 50 % of all pions originate from vector mesons. Predominance of vector
mesons over pseudoscalars is expected from the quark model simply on the basis
of statistical spin factors.
Rp-=~
""----.-----.-----r-----.
0.20
0.15
0.10
t
005 f /\cAc
OL-__~____~II _____~I~
3.5 4 4.5 5
Vs(GeV)
Fig. 4.17
0.6 r- a) + +++++
+
IQ. 0.4
....,.. +
~0.2
a
0.10 b) t+
~+tit t
1<
+
<
a:: 0.05
a 1
3 4 5 6 7 8
Ec.m(GeV)
Fig. 4.18
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 293
The increase of the cross sections coincides with the expected thresholds
for singly charmed baryons (strangeness 0 or 1) around 4.4 GeV. If the in-
crease is indeed due to charmed baryon production, the ratio of charmed bar-
yons to uncharmed baryon production is about the same as the corresponding
ratios for mesons.
As we saw in preceding chapters, the charm hypothesis 106 ,107) can explain
most of the experimental findings connected with the cc-system. The final proof
for charm comes from the detection of mesons consisting of a charm quark and a
light antiquark and their antiparticles. Of course, one expects that according
to the two relative spin orientations each cq state occurs as para (J = 0) and
ortho (5 = 1) particle~ The following nomenclature is generally accepted:
cs-
C +1 : D+ cd, DO -
cu, F+ pseudoscalars
C -1: D cd, DO -cu, F -cs J = 0
The excited states can decay to the ground states by hadronic or electro-
magnetic interaction, e.g. D* -+ D + 1T, D* -+ D +yor F* -+ F + y(see chap. 5.4).
The higher states are therefore expected to have rather large widths. The
ground states (predicted to be the pseudoscalar states), however, can decay
only by weak interaction and hence should be quite narrow. This is because the
strong interaction conserves charm charge and therefore a cq system is stable,
whereas the weak interaction can convert a c quark to light quark.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 295
In order to discuss the weak decays of the D and F mesons we shall very
briefly recall the structure of weak interactions in its simplest form.
Q Leptons Quarks Q
o 2/3
- 1 -1/3
All these doublets contain left-handed particles. The right-handed quarks and
leptons are singlets under SU(2)w. (For larger groups see for example 108 ) and
chap. 6.)
The d and s quarks are eigenstates of the strong interaction. This need
not be true for the weak interaction. Hence the most general possibility is
the Cabibbo structure, which allows mixing between isodoublets. As can be shown
the most general case for 4 quarks is covered by the rotation 108 )
d ' d cos8 + s . s i n8
(5.1 )
SI -d s i n8 + s cos8
The lepton current J~ and the hadron current J h can both be split in a charged
current JC and a neutral current IN with
(5.3)
In (5.5) the electromagnetic current Je~ has been added and the Weinberg angle 111 )
Gw is the SU(2)W x U( 1) mixing angle.We shall not be concerned here with the uni-
fication of electromagnetic and weak interactions but for completeness the
following relations may be quoted:
4G F 2
-~ M - 37.3 GeV
72 - 2M 2 'W- sinGw '
W (5.6)
e/9 = sin GW' MZ = MW/cos GW
where g is the SU(2) coupling constant, and MW' MZ are the masses of the
charged and neutral intermediate boson.
More important for the following discussion is the GIM mechanism 107 ).
The existence of the charm quark had been postulated in order to explain the
absence of strangeness changing neutral currents as in the decays KO ~ ~+~
or K ~ n v v. The hadronic neutral current (5.5) is invariant under rotations
in the Cabibbo angle, as can be seen explicitly. J~ does not contain G and
in particular the terms sinG. cosG (as + sd) are cancelled. Hence transitions
between the d and s quarks are forbidden and the missing of the 16s1 = 1
K decays is thus explained. However, the GIM mechanism automatically forbids
a1so the neutral transiti ons between u and c quarks. The experimental evi-
dence for this expectation will be discussed later.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 297
The charged hadronic current (5.4) contains two parts. One is proportional
to cosG involving the transitions u++d and c++s. The other contains the factor
sinG associated to the transitions u++ sand c++:I. From the hinderance of
strange particle decays relative to non-strange decays one finds sinG ~ 0.2.
Hence.one calls the decays proportional to cosG Cabibbo allowed and those in-
volving sinG Cabibbo forbidden.
Q.
U(IIJ) _s~~L"'.QL s(-h)
cosec"O.97
=lllQI=l
--lllQI=O
d (-I/J) --:Sin9c--- e(IIJ)
Fig. 5.1
The full arrows show the Cabibbo allowed, the broken arrows the Cabibbo for-
bidden decays. Both involve a change of the electric charge by one unit.
It might be mentioned that the c quark was not only invented for the
GIM mechanism but also establishes symmetry bet~leen quarks and leptons. One
consequence is that the sum of all fermion charges is zero (counting the quark
charged three times because of colour) ensuring the absence of anomalies as
required in renorma1izab1e theories.
298 H. SCHOPPER
The most important message we get from this structure of weak inter-
actions for the detection of charmed particles is the strong coupling between
c and s, resulting in a predominance of strange particles in D and F decays.
More specifically one has the diagrams
K- 1];T]', KK Tt
u 5 5 5 U d
u c 5 c U c
D F+ D
Cabibbo allowed (- cose) forbidden (-sine)
where the W boson can couple on the ri ght side to 1eptons (semil epton i c
decays) or quarks (hadronic decays). In both cases DO decays preferentially
into K and F into n. n'. Therefore these strange particles in the final state
are important signatures to observe the D and F.
From (5.4) it can be seen that the Cabibbo allowed decay transforms
c++s whereas c++d is suppressed by tge in the amplitude. Hence one expects
the following decays with leptons only in the final state
(5.7)
Assuming that the decay constants fF and fK are equal, one finds with mF = 2.03 GeV
As will be shown later the decay width into hadrons is of the order
rtot(F):< 10 13 s -1 and hence the branching ratio 8(F-+jJv) ~ 10 -3 to 10 -4 which
is very small. However, the jJ spectrum is characteristic for a two-body decay
and might be observable.
In summary it can be stated that leptonic decays of 0 and F are very weak
and indeed have not been observed so far.
Such decays are of particular interest since they offer the cleanest
way to study the charm current. This is because they originate from the product
J; J h (see 5.2) which implies that the hadronic current appears only once and
is multiplied by the well-known leptonic current. The two possible decay modes
are:
---(' 1+
--<~ 1+
--(' 1+
U C S c IT c
D F+ D
allowed forbidden
The decay rates can be estimated using symmetry arguments and making assumptions
about the form factors l12 ,l14). One finds for example l12 ,l14)
-
r(D +Kev)
+ ~
( mD
5
_+
mK ) cot 2 e . r(K +n e v) (5.9)
+ -0 + 11 -1
'" r( D +K R- v) '" 1,4 x 10 s (5.10)
and with
r(Do+ K*-R-+v) / r(Do+K-R-+v) '" 0.54 (5.11)
one obtains
(5.12)
For the decays of the F meson one can derive similar expressions but they
are less reliable because of nand n' mixing. Estimates are l12 ,l14)
0.9 x 1011 5- 1
'" 1.1 x 1011 s-l (5.13)
0.3 x 1011 s-l
implying
r(F + n'R- ve ) / r(F + n R-v) = 0.29 (5.14)
The decay F + w R- v is OZI suppressed since the w has very little strange
quark content.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 301
For the Cabibbo forbidden decay one finds the estimate l13 )
r(D o -+7fQ,v)/r(D
- 0 -
-+KQ,v)::: 2tg 2 8:::0.1 (5.15)
and similar suppression factors for D+ -+7foQ,+v and F+ -+ KQ,+v. The decay
F+ -+ 7foQ,+v is Cabibbo and OZI suppressed and hence negl igible.
Besides the transition rates one can calculate the momentum spectra of
the leptons. These will be discussed below together with the experimental
results.
5.131 g~~~~_9~~2~~_fQ~_!~Q:~Q9~_~Q~!~Q!Q~1~_9~~~~~
Because of the arguments given above two-body decays of charmed particle
should be particularly suited to be described in terms of simple quark dia-
grams l12 ,114). Amplitudes are usually calculated using colour factors and
exact SU(3) and ignoring gluons. This approach should allow to test the
general structure of weak interactions involving c quarks. Corrections should
be small compared to ratios of decay rates.
302 H. SCHOPPER
u c
K, KilO
d 5
y:,P ...
+ +
--.L.-/a
4>,lJ, ...
S 5
5 C
special since the two diagrams lead to the same final states and therefore
add coherently.
One immediately finds many relations for the Cabibbo allowed decays:
2r (F+~ n~+) = (3/4)2 r(o+ ~ KO~+) =
(5.16)
18r (Do ~ KO ~o) = 9r (F+ ~ K+ KO )
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 303
Decay rates for Cabibbo allowed decays have been calculated taking into
account short distance gluon corrections, SU(3) breaking and form factors l12 ).
The results are given in table 5.1.
colour enhanced
DO + K- p+ 2.3 DO ->- j(*o p+ 3.4
K*- p+ 1.7 KO p+ 3.7
K- 'IT + 1.7 -*0
K 'IT + 2.7
K*- 'IT + 0.96 KO 'IT + 3.7
Besides the graphs shown in fig. 5.2 other graphs can contribute to
two-body decays which are of two types
304 H. SCHOPPER
>--~:
C
F-poTt+
d
Fig. 5.3
The amplitudes for these graphs are multiplied by the factors mn2 / (m
and mK / (mo - mK), respectively,and hence they can be neglected relative
2
m~) r-
to the graphs of fig. 5.2, The expectation
(5.17)
can be considered a significant test of the ideas underlying the quark graphs.
If an additional qq pair is added to the final state of the left graph one
obtains a prediction for the 3-body decay F+ ~ n+n+n-
(5.18)
SU(4) SU(3
L'lC = 0 IL'lcl 1
20
~
84
., (L'l1
8,27
1/2)
(L'l1 = 3/2)
~,
15,
-
15
6
According to the well known L'l1 = 1/2 rule which was found experimentally,
the nonleptonic decays of strange particles (L'lC = 0) are enhanced. From the
above table it is seen that this octet enhancement can be interpreted as 2o-plet
enhancement in an exact SU(4). If this were true the octet enhancement should be
associated to a sextet enhancement of the decays of charmed particles. Whether
this concept is true is one of the interesting questions that can be answered
from a study of nonleptonic decays of charm particles.
for the D decays. The suppression of 84 and hence of 15 forbids the decay
D+ ~ KO~+. As will be discussed below the semileptonic branching ratio for D
decay is 11 3 %,which is not small ,and also the D+'" KO~+ decays have been
observed. In conclusion it seems that there is no sextet predominance and a
different explanation for the octet enhancement has to be found.
Decays of D and F mesons in many hadrons which are Cabibbo allowed are
proportional to cos 4 e and have ~C = ~S = 1. They can be written symbolically
DO ... (K + n 'IT)0
D+ ~ (K + n ~)+ (5.20)
F+ .... (11 + n ~) +, (11' + n ~ )+. (K K + n ~ )+
with C = I, S =0 C = 0, S -1, Q = +1
.... +
~ ~
su, du, ua
is unfavoured for charm particles, but allowed for normal particles like
K*+ ~ K+~+~-.
The inclusive leptonic yield and the total hadronic width can be estimated I14 ,116)
by comparing the graphs
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 307
=(:: r
From the first two graphs one concludes
Assuming that the final state interaction of the quarks indicated by the
bubble in the diagram on the right goes with unit probability one finds
decay as if it were a Do, e.g. to K+n- instead of K-n+. Another way to look for
DO - DO mixing is in associated production of DO, DDN etc., where the kaon
originating from the two charmed particles have opposite charge for no mixing
(S conservation) and equal charge for mixing.
(5.24)
The final state with two equally charged K's arises through both mechanisms,
whereas ee solely through mixing provided the semileptonic ~Q = ~C rule
is valid.
Since DO's decay promptly in-l0 -13 s there is little time for mixing and
hence the theoretical predictions for mixing are very small (much less than
a percent; for literature see l13 ).
~~~ __ ~~E~rj~~~!~1_r~~~1!~_2~_g_9~~~~~
The decisive confirmation for the charm idea introduced to explain the
J/ljiparticle and the other charmonium states is the existence of the charmed
mesons DO, D+ and F+ (see 1.21). Hence it is not surprising that a big
effort was started very soon at SPE.L\R and at DORIS to find these particles.
Initially thi~ search was unsuccessful l18 ) and also the expected rise in the
inclusive K yield at charm threshold was not seen (see 4.41). The experimental
work at SPEAR and DORIS was complementary since MARK 1 concentrated on the
investigation of hadronic decays, whereas PLUTO and DASP were looking for
semileptonic decays. Both kinds of searches became successful in spring 1976
after K identification became possible at MP.RK 1 by time-of-fl ight measure-
ments and electron identification was introduced at DASP by Cerenkov coun-
ters. PLUTO could establish for the first time e-K o correlations.
The observation of hadronic final states has the advantage that in-
variant and recoil masses can be measured. Hence beautiful and very precise
mass values could be published by the SPEAR groups. This is not possible
for the semileptonic decays since the neutrino remains undetected. However,
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 309
The decay 1/1(3.77 GeV) + DO is very close to threshold which allo~ls a very
precise determination of the D masses. The mass is calculated from m = (E2_p2)1/2
where the energy E of the D must equal Eb, the energy of the incident beams.
Eb is known very preci sely and has a spread of only about 1 MeV. Being close to
threshold,p2 ~ 0.08 (GeV/c)2 is small, and thus the effect of any error in p on the
mass is minimized. As a result a mass resolution of about 3 MeV/c 2 can
be obtained, which is better than normal mass resolutions by a factor of
5 to 10 (see chap. 1.13).
The data were collected 104 ) with MARK 1. Charged kaons are identified
by time-of-flight measurements and neutral kaons by measuring the dipion mass
and checking the consistency of the vertex position with the kaon decay time.
t
be of the form P(S) - (1+ a cos 2 8),
lal ~ 1 for any 0 spin and a = -1
for spin O. Fig. 5.6 shows the angular 20
distribution for 0+ and Do decays.
The values of a are a = -1.04 0.10
and -1.00 0.09, respectively, con- 10
sistent with the spin 0 assignment
for the 0 mesons.
o ~~------~------~
In order to verify that a Do 20 ot ... K+ lit lit (b)
is produced together with a Do the
+ -
recoil mass against the K- rr+ system
restricted to the Do mass has been
determined. The result is shown in 10
fig. 5.6~. A clear peak at the mass
of the Do is seen. Thus associated
production is verified which proves O ~~----~ ______~J
the conservation of the new quantum
-1.0 o 1.0
number charm. cos 0
Fi g. 5.6
An important point is that for
o mesons originating from the 3.77 GeV
branching ratios B can be deter-
mined, whereas at higher energies
only a . B can be given.
This requires, however, two assumptions: the ~(3.77) must have a definite
isospin (0 or 1) and its only sUbstantial decay mode is ~O. The results ob-
tained under these assumptions are shown in table 5.2.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 311
D+ -0 +
-> K 7f 1.5 0.6 104
\
modes of the DO and D* to various
0
~l ~
final states 101 ) is shown in
0 fig. 5.7. The results for a . Bare
summarized in table 5.3. As one sees
Ks n:' It" KsTt'TC"
there is not much energy dependence
~ ~1:
~o
~ and the results at 3.77 GeV are con-
~ 20
co
..... firmed. In particular the "exotic"
ia 0 Kt. It; It It- state K-~+~+ is seen again, whereas
~
z
~~
K'!t" It'
80 like K~. The K3~ state is dominated
by Kp~ while the contribution from
LO
II K*~~ is small. If the maxima at 4.03
o 1.8 2.0 1.6 1.8 2.0 and 4.4 GeV in the total cross section
1.6
EFF [[TlVE MASS (StV/e 1)
are attributed entirely to charmed
Fig. 5.7 meson production, then the identified
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 313
Mode E
c.m. (GeV)
hadronic decays add up to only about 10 % of all decays. The reasons for
this small percentage are undetected neutrals and low detection effi-
ciencies. As will be shown below at these energies the D mesons are pre-
dominantly produced via or associated with D* states.
e+e- ~ DO DO
DO 0*0 + Do D*o (5.25)
D*o 0*0
(b)
_0
Ec ,", ' 4 4 15 GeV
........
u
><I>
~
Fig. 5.10
o
<"
Ii The 0* and 0*+ being established
0.
... 100
i II
Ec .. '3 9 -46
as peaks in the recoil mass distri-
::> bution against 0 and 0+ we shall now
Z
turn to the decays of these particles.
Obviously the decays O*O~ OOy and
0*+ ~ O+y should be possible since
they just involve a spin flip of the
c quark (transition between HFS
I~ 2000 2500 states). If the mass difference is
Reco il man (MeV I e') large enough, the strong interaction
decays 0* ~ DOnO, O+n- and
Fig. 5.9 Recoil spectra against D 0*+ ~ O+no, OOn+ should also occur.
316 H. SCHOPPER
These decays are difficult to detect at e+e- energies below 5 GeV since
in this case the pions originating from the 0* have low momenta and with ex-
perimental momentum cut-offs at about 100 MeV/c the detection efficiencies
are low. For this reason measurements in the range E = 5.0 to 7.8 GeV were
performed 121 ). The invariant mass distribution of K ~ shows a clear 0 peak.
If the K combinations in 00 peak are combined with the TIt one expects a 0*
peak in the invariant mass plot. The quantity that is precisely measured is
the mass difference M(Oon) - M(Oo). The distribution of this difference is
shown 89 ) in fig. 5.11. From these data the mass difference 0*+ - 0 is found
to be 145.3 0.5 MeV/c 2 or equivalently
the Q value for the decay 0*+ -+Oon+
is only 5.7 0.5 MeV. Because of this
(al
small phase space the electromagnetic
20 decays 0* -+ Oy can compete with the
strong decays. The mass spectrum of
5 to 7.8 GeV
fig. 5.11 gives also an upper limit on
I, the wi dth of the OM wi th r( OM) < 2 tleV.
The masses of D and 0+ agree well with those determined at 3.77 GeV
(see chap. 5.211).
In fig. 5.11b the nonexotic KTr7T combinations are shown. From the
data at the position of the D* mass the limit 89 )
can be derived which puts a limit on DO - DO mixing (see chap. 5.14). From
the number of events where the K in the DO recoil has the same sign as
in DO a different limit on DO - DO mixing can be obtained 89 ):
t: =- - - - - - - - - (5.28)
where N(K) gives the number of events with kaons of opposite and same sign
respectively. one obtains experimentally t: ~ 0.8.
For complete DO - Do mixing t: = 0 and for no mixing t: = 1. The experiments
are thus compatible with the non-existence of a neutral flavour changing
current.
The data also yield the limits for the Cabibbo forbidden decays
(5.29)
41-1~';-
5 ----!-I---!,/ a> I I I / l:-~I-;I~I~/ (!)
M.ss at poon PI" HIKoJ G,Y HI IG,V MI IG,y
ASSOCI Ale 0 ElECTRON SIGNAl
Fig. 5 . 13
\0
Ko ntlOuum
40
ECftlIifV "
Fig . 5.14
--"
e'p " - pt.~2prongs
v 3.0
>
399 < Eell < S.20 Gpy
Cl MISlDE NTlF IEO HAO"ONS
1- 2.0
- " - HE AV~ lEPTONS
";ii
N
[
. ..
A'
.... OS 10 tS
bl~
'0'0 p. (G pY/c)
Fig . 5 .1 5
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 321
spectra cover similar energy regions and one has to find ways to separate
the two phenomena. This can be done on the basis of two criteria:
I.} The charmed mesons tend to produce final states with more than 2 charged
tracks, w~ereas the heavy lepton has predominantly two tracks in the
final state.
2.} The D mesons decay mainly into K'mesons whereas the heavy lepton has
a very low branching ratio for kaons.
In this way one can distinguish two types of events: high multiplicity,
large kaon content versus low multiplicity, few kaons. These two kinds of
events have been invetigated 127 } quite carefully and it could be shown
that the meson and lepton decays can be separated quite well. The average
number of charged kaons was found to be 0.90 0.18 per multiprong event
and 0.07 0.06 per 2-prong event. This also shows that for charmed par-
ticles (i.e. high multiplicity) the weak current couples strongly to strange-
ness in accordance with the GIM mechanism.
---- 0 -toil<
30 - O-.vK-(892)
20
10
Pe(GeV/c)
Fig. 5.16
322 H. SCHOPPER
by the DELCO detector at SPEAR which covers a much larger solid angle with
C-counters than DASP (60% compared to 7% of 4~). So far the statistical accuracy
that could be obtained does not yet allow such a separation 129 ). Some results
with lower statistical accuracy have recently also be obtained by MARK 1 130)
All these data are in good agreement.
DASP data 126 ) taken at higher Ecm show no drastic change of the spectrum
which is not trivial since they were taken above the thresholds for F and
charmed baryon production.
From the spectrum displayed in fig. 5.16 one can infer a limit on the
leptonic decay D-r ev which would produce a peak near Pe ::: 1 GeV/c. One finds
a(D-r eVe) / a (D-r eX) < 0.09 (90 % Cl) in agreement with the theoretical
expectation (see chap. 5.11). The Cabibbo-suppressed decay D -r ~ e v (see
chap. 5.12) produces a spectrum similar to those from D -r K e v and D -r K* e v
but it is shifted to higher momenta. From the fits to the measured spectrum
it can be excluded that the decay D -r ~ e v is the sole leptonic decay mode.
Integrating over the spectra and comparing to the total cross section
one can derive the leptonic branching ratios. These have already been dis-
cussed in chap. 4.42.
As has been explained in chap. 5.13 the lowest states of the F mesons
tend to decay to final states containing ss which results in nand n'. Since
it is more difficult to detect n's than K's the discovery of the F mesons has to
be considered much more difficult that those of the D mesons which are charac-
terized by kaons in the final state. A first indication of F production was ob-
tained from the inclusive yield of n which shows an abundance at Ecm = 4.4 CeV
(see chap. 4.43). Since n is a frequent byproduct of n' decay, a search for
n includes a search for n' .
e+e - ->- F* F4 nn
(5.26)
ly
~yy
F
low
A good event is characterized by 3 photons, one of which has a low energy, the
other two forming a mass in the n region. In addition an identified charged pion
with momentum above 0.6 GeV/c had to be seen. These events could, however, also
originate from the reaction
e+e - ->-
F (5.27)
4n n
Lyy
324 H. SCHOPPER
[}ASP PRELIMINARY
25
>
'""
M
+
2D
~
~ +
~ 1.5
.,t
+
1.0 15 20 2.5
M~~ (GeV)
Fi g. 5.17
o~~ __~~______~~~
1.0 1.5 2.0
M~~ (GeV)
X2 < 7 IM~~ - M~sl< 250 MeV
Fig. 5.18
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 325
The observed events were fit for both processes. The most recent results are
shown 123 ) in fig. 5.17, where the invariant 7Tn mass is plotted against the
recoil mass. A clear clustering of events can be seen corresponding to (5.26)
or (5.27). In fig. 5.18 the projected n7T mass distribution is shown. The peak
in the mass bin 2.025 to 2.050 GeV is attributed to the F.The events at lower
masses could originate from A2 decays.
The interpretation of the HFS splitting has already been discussed in chap. 2.2.
It agrees astonishingly well with the theoretical expectation.
These data were taken at Ecm = 4.4 GeV where the inclusive n production
shows a prominent peak (see fig. 4.13). However, it would be very interesting
to observe FF production just above its threshold at 4.06 GeV (see table 4.2)
where FF* is not yet possible, in complete analogy with DO production at
3.77 GeV. The right place for such a search seems to be the peak at 4.1 GeV
and indeed a clear n peak has recently been found 123 ) in the yy invariant mass
distribution (fig. 4.15). In this case no low-energy photon was required, of
course. The shaded events in fig. 4.15 show events with an additional electron.
One notices that the n peak is correlated with electrons indicating that a weak
decay \~as involved. No n peak was seen for e+e- energies 4.0 to 4.06 GeV, i.e.
below FF threshold.
326 H. SCHOPPER
Mass r - - - - -21~0:
-- 50
-
MeV F*+tt
yI""'" "'V
2100
2000 n
F+ t.
y \ 2030~50
142.7 MeV \ Y 140.2: 0.9 MeV
!1} n+
1900 \\
DO t+ D+ h
18 63.3 lB5B.~
. 0.9 :!:. 0.9
Fig. 5.19
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 327
Perhaps the most striking success of the charm model is the discovery
of the charmed mesons predicted by the model.The most important results are:
1.) Both the isospin doublet DO, D+ and the isospin singlet with strange-
ness F+ have been seen. As expected the lowest states are the ones
with J = 0 (quark spins opposite). The corresponding states with
J = 1 have also been found and the HFS splitting agrees with the ex-
pectation of the potential model (see chap. 2.22). These ortho states
decay to the ground states by strong or electromagnetic transitions.
The transition rates present no problems to their theoretical inter-
pretation.
The mass spectrum of the charm mesons and the observed transitions
are summarized in fig. 5.19.
Higher excited states, e.g. P states, have not yet been seen.
2.) The weak decays of the ground states of charm mesons are compatible with
the expectations of the minimal model of weak interaction (see chap. 5.1).
The rule lIC = liS = 1 which is impl ied by the minimal theory is drama-
tically confirmed by the dominance of kaons in the final states. Also
the Cabibbo structure of the charm current is evident by the suppression
of lIC = 1, 1I S = 0 decays. Parity violation has been demonstrated.
3. ) The chi ra 1ity structure of the charm current coul d not been verifi ed
yet. The accuracy of the semileptonic decay spectra is not sufficient
to distinguish between V-A and V+A. A distinction between D ~ e v K and
D ~ e v K* would help, since the first decay goes by pure V, whereas
the second contains V and A components. Polarization
measurements would be the most direct way to clarify the
chirality structure. Ao polarization from charmed baryon decays could
offer a possible way.
4.) No sextuplet enhancement is found in charm meson decays. As a conse-
quence, the 1I1 = 1/2 enhancement found in ordinary strange particle
decays cannot be explained by 20-plet enhancement.
5.) No DO - 50 mixing is found. The present accuracy excludes the existence
328 H.SCHOPPER
So far all the observations on charm mesons are compatible with the minimal
theory. This model, however, cannot be exact if embedded in a world with 6 leptons
and more than 4 quarks. Hence it would be very interesting to find deviations.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 329
6. The Upsilon
1jJ'
-34
c::
g -35
<>
:>
c::
.....
~ -36
C>
......
N
~ -37 8 9 10 II
moss (GeV)
From the figures of table 6.1 one can extract the following data which will
be important for the discussion in chap. 6.2.
2 peak 3 peak
r { cryi
r.{
M2 10.06 0.03 10.01 0.04 GeV
r"i B dOl
dy y=O
--- 0.011 0.007 pb
x2 / DF 19.3 / 18 14.2 / 16
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 331
2 peaks 3 peaks
~~!_~~~i!~!iQ~_~~~r9i~~
Even before the detection of T, the standard model with a linear
confining potential had been extended 133 ) to larger quark masses. The
excitation energies above the lowest-lying state are shown in fig. 6.3
as functions of the quark mass. The
~-,~------------------~ I OOO
essential point is that for a linear
potential the excitation energy is re-
duced proportional to mQ- 1/ 3 Since the
600 continuum boundary on the other hand goes
up one expects more bound states. For
--"!--~400 mQ ~ 4.6 GeV not only the IS and 2S states
are bound but also the 3S level.
200 zoc
Most of the theoretical estimates are in favour of 1/3 (for a comparison see
Ref. 132), but the uncertainties are so large that no definite conclusion can
be drawn.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 333
where one finds two jets with distinct lab energies (~ 5 GeV) accompanied
by a monochromatic photon. The angular distribution allows us to differentiate
between jets originating from gluons and from quarks.
The existence of the charm quark made it possible to group the 4 quarks
and 4 leptons into weak SU(2)W doublets and a minimal theory with GIM mechanism
could be formulated for the weak interaction. It is compatible with
334 H. SCHOPPER
all experiments (see chap. 5.1), except for the non-observation of parity
violation in atomic transitions 14o ) which, however, needs confirmation
and the small CP violation in K decays. The beauty of this
model is the symmetry between quarks and leptons which is needed
to cancel the triangle anomalies which would otherwise have destroyed
the renormalizability of the weak and electromagnetic gauge theory141).
A very important point is, however, that the 6-quark model gives CP
violation in a natural way. This is achieved by a generalization of the
Cabibbo structure of the hadronic current (see chap. 5.1). The most general
way of mixing the lefthanded charge 1/3 quarks is143)
d' = c 1 . d + sl c3 s + sl s3 b
io io ) . b
b' sl s 2 . d - (c 1s 2c3 + c2s 3 e ) . s + (c l s 2c3 + c2c3 e
(6.3)
whereas the 2/3 charge quarks u, c and t remain unchanged. Here the
abbreviated notation sl = sin 01' c l = cos 01' s2 = sin 02' '" has been
used and 01 is identical to the Cabibbo angle. The 4-quark model is repro-
duced by 02 = 03 = 0 = O.
PROPERTIES OF CHARMONIUM AND CHARM PARTICLES 335
Because of the limits for 82 and 83 one might speculate if they are
exactly zero. This has quite amusing consequences. The decay of the strange
particles to ordinary hadrons is only possible because of the d-s mixing.
Without this "feed-through" the lightest strange particles would be abso-
lutely stable. With 6 2 = 6 3 = 0, b would not mix with d and s, as can
be seen from (6.3) and as a consequence particles like bq would be stable
if the b quark is lighter than the t quark or v.v. Such stable particles
of a new kind of matter could be captured in nuclei and the consequences
for the ensuing X-ray spectra and formation of hypernuclei have been recently
discussed in detail I44 ).
It is hoped that the e+e- storage rings PETRA, CESR and PEP which will
come into operation soon might make it possible to shed some light on
these exciting and fundamental questions.
336 H. SCHOPPER
Appendix 1
References
DISCUSSION No.1
SIZZARRI:
My question is on the F*-F mass difference. You
said it is determined mainly by the ~ energy which is
given as 120 40 MeV, but you have a cut at 140 MeV with
a resolution of 60 MeV. Since this cut presumably excl-
udes a large number of events I wonder if the observed
peak at 120 MeV could not be artificially produced.
SCHOPPER:
The energies of the ~ rays are determined not by
direct measurement, but by kinematical (2c) fits to the
events. The energy resolution obtained in the fit is
much narrower than given by the direct measurement. The
large spread of 40 MeV quoted in the ~ ray energy comes
not from the errors on the individual fits, but from the
uncertainty in the production hypothesis, i.e. whether
it is:
345
346 DISCUSSION
SMITH:
)'c
What fraction of the F F -+ F y F events are lost by
the cut on '( ray energy?
SCHOPPER:
WU:
In the SPEAR result there is a peak in R at about
4.43 GeV. I understand that your FF* events come from
a scan from 4.32 to 4.44 GeV. Why do you choose to run
on the left side of the peak? What exact beam energies
do these four FF* events correspond to?
SCHOPPER:
WU :
SCHOPPER:
DEO:
MARTIN:
It shall be covered in my lectures.
CELMASTER:
I have a comment to make about potential models and
various other things you said this morning and I also
have a question afterwards.
There is one thing that can be observed from all
these models. By and large, most details about charm-
onium were postdicted and not predicted. Correct pred-
ictions are common to most of these models, where these
models differ are in the "postdictions", to my view this
is ugly and I don't think that it is very valuable to
overly refine these models. My bet is that they will
have minor failures for "bottomonium". Another thing to
consider is the following:
All predictions are based on the Born approximation.
In particular this holds for the "Coulomb-perturbation"
prediction 3'P. _JR
& t -=.8
"'?1 -~"P
0
If you solve the Schrodinger equation directly for those
three states, that is, by including the spin orbit etc.
forces in the potential, it turns out that
3 ~"'D
1; - r1 =.5
3'P. _ l'P,
1 D
Perhaps this shouldn't be taken too seriously, as other
types of perturbative effects might compete with this.
However, we do see that it is possible, within the exist-
ing framework to explain such small effects; it is not
necessary to introduce long range scalar-scalar inter-
actions etc. Finally, regarding ~c; I think that it is
not wise to try too hard to explain its mass. There are,
as Professor Lipkin has said, two other difficulties
which no one can explain in the potential model context,
that is, its very small hadronic width and the low mag-
netic transition from J/~ smaller by a factor of 15 from
prediction. This leads me to believe that X (2.8) is
not understandable in potential models, unless it is ~
the 'tIc
348 DISCUSSION
CELMASTER:
My question is: can you repeat why you use for the
F signature the'~data in the way described during your
lecture?
SCHOPPER:
LIPKIN:
SCHOPPER:
CARTER:
MARTIN:
can be changed.
OLSEN:
The way I understood the raw~J spectra that you
showed this morning, was that they corresponded to events
which had already passed the ~~ , (one less than 140 MeV)
2 charged particles. Is that correct?
SCHOPPER:
Yes.
OLSEN:
SCHOPPER:
The improvement comes from the requirement of a
charged it in one of the spectrometer arms.
OLSEN:
What is the momentum acceptance of these spectro-
meter arms?
SCHOPPER:
Please reserve the question for Bjorn Wiik.
OLSEN:
,
Are all the n. $ observed consistent with having
originated from F's?
350 DISCUSSION
SCHOPPER:
LIPKIN:
CABIBBO:
SCHOPPER:
CABIBBO:
PREPARATA:
The value (:r-+ cpTtlt /"J-'f C01111 ') ~ ~ which you showed this
morning to remain after subtraction of known resonances
is in good agreement with the following idea. The 1r~
system is in S-wave and for the 0+ I=O resonances the
strange - nonstrange quark admixture is, as experiment-
ally suggested, non-negligible.
LIPKIN:
SCHOPPER:
WEEKS:
LIPKIN:
~1(
Note that there are no gluon intermediate states.
SCHOPFER:
LIPKIN:
LIPKIN:
WETZEL:
LIPKIN:
CABIBBO:
PREPARATA:
LIPKIN:
~,t'
1L
These diagrams can be calculated for on-shell KK states,
and the results agree with experiment. Tne corresponding
proces2 J 1,1.
't' -
-
DO - O,U),u:l, ... ~ plt, '71.,... is not possible
I"
\1=:."+
PREPARATA:
If we want the OZI rule to be good for the J/~ decay
the only way to keep the offending diagram
1)
~ D
from destroying it is to assume that ~l~~DD)~lJ'I'\l~t>'5)
i.e. a very big violation of SU(4). This is borne out
by calculation in the Geometrodynamical approach as I
have shown last year at the Trieste Topical Conference
on Particle Physics.
LIPKIN:
ROLANDI:
o
In reference to the recoil mass spectrum from D ,
I cannot understand why the peak "on the right" from
SCHOPPER:
The reaction takes place near to the threshold so
that the recoil momenta involved are all very low, as is
the Q-value of the rl! -. bltC) decay. At higher energies
the peak is significantly wider due to the smearing
effect you are thinking of.
RECENT RESULTS FROM DASP
B.H. wiik
Hamburg, Germany
357
358 B. H. WIIK
Fe 5 OK SR
(J" (nb)
15r-------,r------,-----,
10
?0.3<p<1SGeVIc
+aamomenta
O~--~I~--~~--~
3 4 5 6
Vs(GeV)
10 _ I
I
~ vslGeV)
l-
t- o 3.6
-
x 4.05
N
~
<!)
r:-
-
:0
c:
blo-
'0'0
t-
t- t
wf} t-
0.1 ::-
1j j t \ 1
1
0.01 t
.L-....1----L_L---1-_.l.--.I....1----L.---l_..I....-......I.---I
0.5 1.0 1.5
E I GeV)
Fig. 3 The invariant cross-sections E/(4rrp2) dcr/dp for
the sum of K+ and K- pro duct~on
. at c.m. ener-
gies of 3.6 GeV and 4.05 GeV.
RECENT RESULTS FROM DASP 363
I
3 -
(a)
2
R"ew
~~~ ~++ +4 +
~ f~+
+-
1 r-
<} <}
-
o .6.
1T
-1 RKew
-2 -
I I
4 Rnew
1
-
3 - -
(b)
~-
r-
~~ ~
lr- ~~ \ . +*++H + -
9 .:.
or-~A
1- +R~ew -
~ R"ew,no H.L.
2r- I I
I
3.5 4.0 4.5 5.0
Vs (GeV)
Results in agreement with these data have been obtained for neu-
6) 7)
tral kaons by the PLUTO group and the SLAC-LRL group
The charm model predicts that there should exist mesons with
+ -
both charm and strangeness. The lowest member of this set F (cs)
should decay predominantly into an ss system leading to final states
containing KK, . n or n'. A step in the n yield would. therefore.
be a signature for pair production of F mesons. Note that an in-
clusive n search also includes a search for n' since n' + nrrrr oc-
curs with a branching ratio of 68%.
, e + e - annihilation near D* thresh-
Experience has shown that ~n
-* and D*-*
old DD D production is copious. In analogy to this we as-
-* and F*-*
sume FF F production to be enhanced. The first excited
state F * is ,
pred~cted to be close to the F such that F* + yF or
F+ + rrDF (electromagnetic only. strong forbidden by isospin) are
the favoured decay channels. To enhance the signal we searched 8 )
for final states with an n and a low-energy photon. We identify
the n by its decay into two photons.
than 1.2 GeV/c. All such pairs of photons are used to reconstruct
the m mass. The events were divided into two classes depending
yy
on whether the event contained a photon with energy less than
0.14 GeV or not. The myy distribution of events containing a
low energy photon is plotted in Fig. 5 for all energies except
4.4 GeV. and in Fig. 6 for 4.4 GeV. The solid line represents a
smooth estimate of the uncorrelated background obtained by forming
the invariant mass of photons from different events having the
same number of photons. The background spectrum is normalized to
the data by requiring all combinations with m > 0.7 GeV to be
yy
ascribed to background. There is a clear n signal at 4.4 GeV, not
observed at other energies. To ascertain that the n is correlated
with a low-energy photon, we also evaluate the m distribution
yy
for events without a low-energy photon and plot the ratio of the
m distribution for events without a low-energy photon. This ra-
yy
tio is plotted in Fig. 7 as a function of myy At 4.4 GeV there
is an enhancement at the mass of the n showing that the n is in-
deed correlated with a low-energy photon.
4.0 GeV
60
40
4.15 GeV
>~
-
~
o
LIl
I II
c
.2 4.5 GeV
g 60
:0
~ 40
-...
u
o
~
.c
E
J
Z
4,4 GeV
80
>CI)
~
0
Lf')
........
Ul
C
0
.-
0
40
C
..0
E
0
u
20
m (yy) in GeV
~} l.! III
.4
.3
.2
f II
.1
~.1!!! III
.1.
.3
.2 I If
-9.4 ~
.1
I f I I II
;>-
0
c: .3
.......
I II
-
~
~.2
.1
4.4 GeV I
- .3 .4
0
0
Ifj!!IlllJ
0:::
.2
4.5 GeV
.1
.4
.3
.2
I I Iff f r I I J
5.1 GeV
.1
0.2 0.4 0.6 0.8
M (yy) in GeV
Fig. 7 Ratio of the number of combinations for events
having a low-energy photon to the number of com-
binations for events not having a low-energy
photon as a function of m
yy
RECENT RESULTS FROM DASP 369
(1)
(2)
These are 2C fits because of the mass constraint m and the re-
yy
quirement that for (1) TIn and the missing vector must have the
+
same mass ~; for (2) the Yl F- system and the missing vector
ow
must have the same mass mF*. A total of 15 events satisfied the
criteria with a X2 of less than 8. The same events also fit the
hypothesis (2).
..
~ 0.15
u
">GI 0.10
--
C> 0.05 ~
-
t:!
F 0.30
1..0-5.2 GeV
~
0.25 ~
excluding 1..1. GeV .
I
~
~
0.20
0.15
0.10
0.05
0
1.0 1.2 1.1. 1.6 1.8 2.0 2.2
M (lilt) Fitted (GeV/c 2)
60
Any No. of Y' s
wm No. Y
-
III
c
~
ell
40
Z
20
2 4 6 8 10 12
N prong with e
Q5 1.0 1.5
bl cf
'tJ 'tJ Pe (GeV Ie)
lepton contribution was estimated for a lepton of mass 1.9 GeV and
a branching ratio of 30% into multiprong final states.
.
...---,
- >
u
ell
(!)
4.0
0) e+e- - e:!: + ~ 2 prongs
3.99 < ECM < 4.08 GeV
.0
C
'--'
o +ev K
3.0 o -evK (892)
III
0'1
C
0
L-
a. 2.0
N
1\1
+ 1.0
+,
ell
,+ 0.0
ell
+
ell
b
'0 l '0a... 0.0 0.5 1.0 1.5 2.0
Pe (GeV/c)
+ -
Here a(e e + C.C.) denotes the effective cross-section for produ-
1 ]
cing the lightest charmed hadron stable against strong and electro-
magnetic decays. These particles might eigher be produced directly
or result from the cascade decay of excited charmed hadrons. The
cross-section a(e+e- + C.C.) was obtained by subtracting the cross-
1 ]
section for "old" hadron production (R = 2.5) and the cross-section
for heavy lepton production (m = 1.91 GeV) from the total cross-
section.
RECENT RESULTS FROM DASP 377
I I
2.0 f-
- 1.5 f-
f
-
-
r
.0
c:
b 1.0 f-
05 f-
g o. 'f
i
c.!) 0.25 f- 0'" (e+e--- e +~ 2 prongs)
z
b) 2 (J (charm)
I
~
u
~ 0.20 f-
a::
III
+++t+ I
u
Z 0.15 f-
....
0
n.
UJ
:::! 0.10 f-
~
UJ
VI
UJ
c.!)
005
c(
a::
UJ
~O.O
3.6 3.8 4.0 4.2 4.4 4.6 4.8 5.0 5.2
ECM (GeV)
B R (C -+ e T )() = O. .f 4 o. 03
obtained by averaging over all energies between threshold and
5.2 GeV. These values were extracted using PLUTO 3) data for the
total cross-section and the error quoted is mainly systematic.
Evaluating the branching ratio using the SPEAR data 2) on the to-
tal cross-section as input lead to an average semileptonic branch-
ing ratio of 0.08 0.03.
The results listed below are not yet final. The data on
reaction (1) are discussed above. Data on reaction (2) were ob-
tained by selecting events with a muon in one of the spectrometer
arms and a non-showering track observed either in the inner de-
tector or 1n one of the spectrometer arms. A charged particle was
called a muon if it had a momentum greater than 0.7 GeV/c, gave no
signal in the threshold Cerenkov counter, suffered an energy loss
consistent with that of a minimum ionizing particle in the shower
counter and penetrated at least 40 cm of iron for momenta between
0.7 GeV/c and 0.9 GeV/c and 60 cm of iron for momenta above 0.9
GeV/c. Muons between 0.7 GeV/c and 0.9 GeV/c were required to hit
the range chamber within 15 cm of the projected track direction.
A total of 14 events satisfied these conditions.
The observed muon yields were then corrected for the follow-
ing contributions:
~~ ( nb/GeV )
0.5
0.25
of- - +
Fig. 13 The lepton spectrum measured in e e ~ e(~) + T-.
The electron and muon data are combined for momenta
above 0.7 GeV/c.
RECENT RESULTS FROM DASP 381
We find
or
Be =0.2:!: 0.03
1 3)
This is in agreement with preV10US measurements and with theore-
14)
tical predictions .
~ p assuming CVC.
382 B. H. WIIK
vl
"t-
A
e - . ....-
V. .V Il
t-
B
a a
d s
-cos9c -Sin9c
c ___"t.-___ W:: ~-
~~
9 gf~cosec
--------------- ==
1
.f. 04 'me = 4
1M!
"t
I
Using this as an input we expect to observe N(en) events with N(en)
= 2.55 B2.
e
With B
e
= 0.2 this yields 10.2 events. This is a firm
prediction of the theory; however, note that it depends strongly
on B .
e
The decay T- + V p- + V ~-TIo, where both photons from the
T T
decay are not measured, populate the en class. With B = 0.24 we
p
expect 1.75 en events from this source. There ~s a further 2%
chance to mislabel a hadron in the inner detector as an electron.
From the number of observed two-prong events we estimate 0.25 events
from hadron misidentification. Subtracting the two background events
(neglecting the charm contribution) leads to B B (0.004 0.005)
e n
compared to B B
e n
= 0.02 expected.
Alt
= .f-. __ '\J
2. A f"l
RECENT RESULTS FROM DASP 385
DASP
PRELIMINARV
e+ e - -- + 0
Tt- Tt + lch
~ : ch = electron
10
all
5
a) 0 OL---'---"""10.f""'5~~~~""'1.0~---L----'1.5
o ~ ___ ~~~~~~~~L-_~
+
Fig. 15 a) The distribution of M(n-no) observed for events
with the topology e+e- ~ nno + T+.
+
b) The M(n-no) distribution for events of the same
topology as above but with p n-n
+
> 1 GeV/c.
RECENT RESULTS FROM DASP 387
DASP
PRELIMINARY
10
~ ch = electron
--~---
/
I
I
....
o /
o 1.0 2.0
+
Fig. 16 The momentum distribution of the (n-no) system
observed in e+e- + nno T~. The mass of the
nno system was between 0.5 GeV and 1.0 GeV.
388 B. H.WIIK
+
These events are plotted versus the mass of the n-n o system
in Fig. l5a. Events with an identified electron are hatched.
Events with the p band (0.5 GeV < M + 0 < 1.0 GeV) are plotted
n-n
versus the momentum of the nn o system ln Fig. 16. The momentum
distribution expected from the decay T ~ V P is shown as the dot-
T
ted line. Note the flat distribution above 0.9 GeV/c ln agreement
with the expectation. The enhancement at low momentum is pre-
sumably caused by multihadron events. To reduce the background,
+ 0
only events with a (n-n ) momentum above 1.0 GeV/c are considered.
The nn o mass distribution for these events is plotted in Fig. 15.
From these events we determine a branching ratio B
e
B
p
= 0.048
0.018.
With B
e
=
0.2 we obtain B
p
0.24 0.09 ln good agreement
with theoretical predictions.
I
I
B
jJ
IB e = 0.8 0.3
e = jJ =
B B 0.20 0.03 I
T ~ vK-+
~ '" 0.07 0.06
T ~ vn
B
e
. B = 0.004 0.005
(T ~ vn not es tab lished)
B
p
= 0.24 0.09
BK < 0.016
RECENT RESULTS FROM DASP 389
REFERENCES
11) A. Ali and T.C. Yang, Phys. Letters 65B (1977) 275 and
private communications.
I. Hinchliffe and C.H. Llewellyn-Smith, Nuclear Phys. Bl14
(1976) 45.
F. Bletzacher, H.T. Nieh and A. Soni, Phys. Rev. l6D (1977)
732.
K. Kajantie, Phys. Rev. Letters 65B (1976) 69.
V. Barger, T. Gottschalk and R.J.N. Phillips, Univ. of
Wisconsin Report.
M. Gronau, C.H. Llewellyn-Smith, T.F. Walsh, S. Wolfram and
T.C. Yang, Nuclear Phys. B123 (1977) 47.
12) M.K. Gaillard, B.W. Lee and J.L. Rosner, Rev. Mod. Phys. 47
(1975) 277.
390 B. H. WIIK
ATWOOD:
You mentioned the existence of a dip in the inclu-
sive e+e-~ X cross-section distribution at Ecm 4.3
GeV. How deep is this dip and does it give any trouble
to the existence of heavy leptons?
WIIK:
I did not mention any dip in the inclusive electron
cross-section. We do not have sufficient data in that
region to decide that. The total cross-section e+~~
hadrons does indeed have a dip in this region. However
note that a heavy lepton with mass 1.91 GeV contributes
less than one unit of R at 4.3 GeV and hence it might
be difficult to get conclusive evidence for or against
a heavy lepton from such data.
ATWOOD:
391
392 DISCUSSION
WIIK:
SCHOPPER:
SANDA:
WIIK:
WHITE:
WIIK:
A. Martin
Geneva, Switzerland
1. INTRODUCTION
One of the most striking events in the last three years has been
the discovery of the new particles. By new particles I mean the J/~,
~/, X's and also the charmed bosons. There are other new particles
such as the heavy lepton T or the "baryonium states" or the candidate
for charmed baryon that I shall leave aside. The reasons I became
interested in this particular group of particles are:
ii) that it appears that the members of the J/~ family can be des-
cribed as a non-relativistic system of a charmed quark-
antiquark pair with great success (of course, the non-
relativistic Schrodinger equation can result from the
reduction of a relativistic equation), with, of course, a few
exceptions; I realized that this is a sufficiently simple
situation to allow investigation by rigorous methods, within
395
396 A. MARTIN
Table 1
Photon Stable
~ j""C<O"
Neutrino Stable
Stable
-6
S Meson].l ].I "'e +v+v 2,2 x 10
(ou V + v?)
+ +
11 "'].1 + V 2,5 X 10- 8
Mesons 11 11- ... ].1- + V
11 ... 2y T < 10- 15 (7)
11 + 11 + 11 ;
.' {
11 + 11; -8
].I + V; 1,3 0,3 x 10
+
].1- + y + v; ].I + rro + v;
Mesons K e + y + v; e + rro + v.
KO = KO + KO ... rr+ + rr- 0,85 X 10- 10
I 12
P stable
Nucleons N ... P + e- + V (ou v?) 1110 220
B
A
R
Y
o 1:;+ ... {' P + rro 0,8 X 10- 10
N N + 11-
Hyperons
S 1:;- ... N + rr- 0,8 x 10- 10
1:;"'Ao+y very brief
-10
1 x 10
NEW PARTICLES 397
The J/~ and ~' are very narrow resonances produced in e+e- col-
lisions at 3095 and 3685 MeV, respectively. Their total width is
estimated by the following method: one assumes that the cross-
section has a Breit-Wigner form. Integrating over energies after
subtraction of the background one finds the production width which
is r e , independently of the energy resolution of the machine:
J~~+p--(t:')dE'
Jo-T (')dE'
re = r r tot
II
We now want to know the quantum numbers of the J!t/J and t/J'. In
both cases it is possible to show that the resonant amplitude in
e+e- + ll+ll- interferes with the Bhabha term, destructively below
the resonance energy, constructively above. This shows that J!t/J
PC
and t/J' have the same quantum numbers as the photon J = 1
and K-K+*. Data indicate that the maximum amount of octet compon-
ent is in amplitude 12%.
c
3iT's
One has singlet and triplet states depending on how one com-
bines the spin of the quarks. The charge conjugation is given by
c= (_,,)1-+5
parity is given by
L..
p =-(-1)
1..'=2.,
J... =0 l =1 __..-""'--___--,
.....-"
r-----~:r.
-------"-
.. - - - -
_~(
___
0
~
~") '3F-
......""""-::--_ _
2+1'
,-+
"!'1--c=-:----- ----~
1+1 /+-
----OL.+7:.,.::-
402 A. MARTIN
In (i) and (iii), decays of the ~' to three states + one pho-
ton have been observed, the energies of the states being 3.41 GeV,
3.50 GeV, and 3.55 GeV. In (ii) an additional state was found. In
fact, there was an ambiguity to resolve: in the decay ~' + y + X,
X + Y+ J/~, which y is emitted first? This has been answered be-
cause as X is in motion, the second y is not monoenergetic. The
states found are 3.41, 3.45, 3.50, and 3.55 GeV. The 3.45 state,
considered as doubtful for some time, partly because of low statis-
tics, partly because of its absence of decay into hadrons, has re-
cently been confirmed by PLUTO.
The I 3410 and 3550 states I are seen to decay into n+n- and
K+K- and therefore they have I Jeven, P = C = +1 I.
The state 3510 does not decay into n+n- or K+K-. This is not
..
a very severe restr1ct10n. . excludes 0++
As f ar as I can see, 1t
++ ++
2 4 etc., but no more.
Hence
0
-+ ~ '34$0
or
spin-O mesons
The second procedure, when it can be applied (i.e. for 3414 and
3552) is the most advantageous one because it minimizes the number
of amplitudes since the final particles have spin-O. This is what
has been pointed out by Kabir and Hey in particular.
where ey and eM are the angles of the y and the mesons with the
beam and ~M is the relativ~ azimuthal angle. These are complicated
expressions depending on J+l parameters which are the J+l amplitudes
for producing a X with helicity 0, 1, . , J. Irrespective of the
value of these parameters the spin is fixed by the (perfect!)
knowledge of W2.
406 A. MARTIN
if JX =: 0 At'- :::' 1
JX := 1 At" := - 3/r:
1X 2 Ap :: - i/1
Now we come to another X state which lies below the J/~ and
has been observed at DORIS 4 ). In the e+e- + 3y spectrum a peak
is observed at 2850 MeV in the yy system after subtraction of the
contributions of ny and n'y. Though this state has not been ob-
served at SLAC, evidence for its existence is accumulating at
DESY. Naturally, the classical arguments of Yang apply to this
state which decays like X +yy, and therefore it has J ~ 1, and
C = +1. It is tempting to attribute the quantum numbers 0-+ to
this state, but not yet fully justified.
New states have been observed around 1870 GeV. In fact, they
have been looked for very actively before they have been discovered,
for two reasons.
408 A. MARTIN
B S Q y C
AQ: A C. a - A5
is strictly forbidden.
So one expects
while
is suppressed, and
ii) At 4.03 GeV e+e- energy one observes peaks in the recoil spec-
trum of the DO and D from which one can infer the existence of
DO*, mass 2003 MeV, width < 40 MeV, with decay DO* + TIoDo and yDO
and D*+ at 2010 MeV, with D+* + TI+Do. There are other peaks but
it has been shown that these can be accounted for by kinematical
reflections and are caused by the very small Q value of D* + TID.
Decays into n*D* have also been observed.
Now one has to attribute quantum numbers. The small mass dif-
ference DOD+ makes us put it in an isospin doublet. The spin is
tentatively taken to be zero because a spin-l would, in general,
show decay correlations which are not seen.
The decay D*o + y + DO shows that not both DO and D*o have
spin-O. Spin-O is favoured for DO by the angular distribution in
e+e- + 3.772 + DOnO 15), hence we are tempted to attribute spin-l
to D*o. If, according to the theoretical prejudice, DO is 0-, D*o
is 1- because of the existence of the strong decay D* + D + TI.
Pushing the analogy with the 6-T puzzle, we can ask ourselves
if there is mixing of the DOno as in the KOKo system. More exactly,
is the mixing so fast that produced DO,s decay in K+n- as well as in
K-n+? The test consists in looking at the decay products of e+e- ~
~ DOno. In the absence of mixing we should observe only K+K-n+n-.
For 100% mixing we should observe e+e- ~ DOno, where D is the short-
s s s
lived combination (notice that e+e- ~ one virtual photon ~ DODo is
s L
forbidden by CP conservation), and events with final states KKn+n+
should be as frequent as K+K-n+n-. Experiment by the SLAC-LBL group
indicates that the mixing is less than 16%.
Once you agree that the "new particles" (1975-76) are made of
a quark-antiquark pair, you have to decide what is responsible for
the binding. There you can start at various levels, from a funda-
mental point of view to a purely phenomenological point of view.
The fashionable point of view consists in saying that quarks are
bound by coloured gluons, but all you can do is to compute the
single gluon exchange which produces a kind of relativistic Coulomb
potential in which you can even decide that the coupling constant
dies logarithmically for r + 0 because of asymptotic freedom;
here at least one has a relatively clear prescription for the
choice of the vector character of the potential, and this is rele-
vant when one goes to the non-relativistic reduction to know what
will be the spin effects accompanying the "Coulomb" force. However,
NEW PARTICLES 413
nobody knows how to compute the "confining" force, its spin depend-
ent effects, etc. So, depending on what you assume in your rela-
tivistic equation as interaction, you get different results in its
~
o 3P
3-~
-i --
3D
( 2-- l.- r
25 ( ,0
- ....
r-
2++
~
1 +~
I
i +-
ott
,,--
"'5
~ 0
-+
.: (_-i)I..+S
414 A. MARTIN
~f
What Harald Grosse and I have proved 22,23).~s .
th~s: ~
Theorem I
If
and
(#I. )
and as well
Theorem II
-
d
Il.r r
we get
For those who think that these two theorems are not very trans-
parent let me indicate that they tell you that in
v::-- + 'B + Cr
-J
_I
o
f(oC)rllidol f(O<) >0
In particular you can weaken the "Coulomb force" if you like to have
asymptotic freedom, and take, for instance, instead of -(a /r)
s
" Sr
2..
~ u.1 elr
ei,.r
jV",ld,r
The decay width of the ~' is 2.1 keY, while that of the J/~ is
4.8 keY. What potential will produce such a ratio? In a harmonic
oscillator potential the wave function at the origin of the 2S state
is larger than that of the IS state. This is a clear indication
that a pure harmonic oscillator is not acceptable. In a purely
linear potential the magnitude of the wave function at the origin
is the same for all = 0 levels 2") ,for it is given by
00
Theorem III
-:3 " ~
r + B + CJ'"
is concave and
One problem about lepton width is that of the ~", the 3.77
resonance predicted long ago as being a D state, and recently dis-
covered.
We have already spoken about the 3.77 GeV tjJlI state which is
slightly above the charmed particle production threshold. The
description of this state as a pure quark-antiquark pair is still
reasonably good. For higher states it may be questionable in the
sense that a resonance above the charm threshold can be made of the
physical charmed and anticharmed particles. This is what is called
by De Rujula, Georgi and Glashow "molecular charmonium,,27) and
which, as pointed out by David Jackson, is not very different from
thinking that the p meson is an L = 1 pion-pion bound state. On
the other hand, it seems illogical to disregard completely the se-
quence of higher states produced by the quark-antiquark potential
which has been so successful in the description of the lower states.
h t h e eX1st1ng
W1t .. . 1 an L -- 0 J PC = 1
potent1a . pred 1cte
state 1S . d
around 4.14 GeV. This is a bit too high for the prominent peak at
E = 4.028 GeV which has been extensively used as a source of D's and
D*.
However, it seems to me that the most reasonable picture which
has already been proposed some time ago on a pure theoretical basis
.. 28).
by Dashen, Healy and Muz1n1ch 1S that the confined quark-
antiquark channel is coupled to the physical particle channel. De-
pending on the energy these channels are closed (and then their
contribution to the wave function is very small) or open. In this
way the two pictures are reconciled and the only risk is that there
NEW PARTICLES 421
REFERENCES
12) G. Karl, S. Meshkov and J.L. Rosner, Phys. Rev. D 12, 1203
(1976).
P .K. Kabir and A.J .G. Hey, Phys . Rev. D 13, 3161(1976).
H.B. Thacker and P. Hoyer, Nuclear Phys. B106, 147 (1976).
13) This results from a discussion between P.K. Kabir and the
author.
24) See, for instance: J.D. Jackson, Lectures at the SLAC Summer
Institute (1976).
27) L.B. Okun and M. Vo1oshin, Z. Exsp. Teor. Fiz. 23, 369 (1976).
A. De Ruju1a et a1., Phys. Rev. Letters~, 317--(1977).
28) R.F. Dashen, J.B. Healey and I.J. Muzinich, Phys. Rev. D ~,
2773 (1976); Ann. Phys. 102, 1 (1976).
29) K. Lane and E. Eichten, Phys. Rev. Letters 12, 477 (1976).
DISCUSSION No.1
LIPKIN:
MARTIN:
SCHOPPER:
ROLANDI:
425
426 DISCUSSION
MARTIN:
ZICHICHI:
SCHOPPER:
BALDINI:
MARTIN:
BIZZARRI:
MARTIN:
CELMASTER:
LIPKIN:
WETZEL:
MARTIN:
BERLAD:
V<rj VCr)
r
A
and not the Coulomb-like potential (as in B) of the
charmonium calculations, where one obtains:
V(Y') rv
-r.. o
...1
T
MARTIN:
DISCUSSION No.2
WETZEL:
MARTIN:
DINE:
WIGHTMAN:
SCHECHTER:
MARTIN:
CELMASTER:
MARINESCU:
MARTIN:
MOTTOLA:
MARTIN:
MIETTINEN:
MARTIN:
MARINESCU:
BERLAD:
MARTIN:
MIETTINEN:
LIPKIN:
BACE:
MARTIN:
= o (normalization)
Hin Chen
l1assachusetts Institute of Technology
Cambridge, Mass. USA
EP Division, CEP~, 1211 Geneva 23, Switzerland
INTRODUCTION
435
436 M. CHEN
3) One can vary the type of beams and targets in the experiments
to have a better understanding of the nature of the production
mechanism.
4) The mass spectrum in the final state covers wide range and
therefore is suitable for searching for new particles with
unknown masses.
5) One can study the correlations between the new particle with
the rest hadrons.
P+ nucleus +J + anything
+ - +
4. e + e or)l +)l
oII
>-
x Cord en et 01.
bl>- Kourkoumelis et 01.
"C"C
t o Amoldi et at.
181 Busser et 01.
o Snyder et 01.
Anderson et 01.
<> Antipov et 01.
l;. Aubert et 01.
VS in GeV
iii) At the lowest ISR energies IS = 23.S and IS = 30.6 GeV the
cross-sections are in good agreement with the FNAL measure-
ments at comparable c.m. energies. Even at such high energies
the J production cross-section continues to increase with
(4)
energy ,although, at a somewhat slower rate :
do
dy y = 0
Vs 62.4)
5.S
+ 5.8
- 1.8
do (IS 23.4)
dy y = 0
p + A -+ J + D + D + X
L
~
f.l
+
f.l l- f.l + X'
o( J DD )
~ 0.0 1
o( J )
NEW PARTICLE PRODUCTION IN HADRONIC INTERACTIONS 439
E A e - B PT (1 - x ) n
dxdp 2
t
B 2.05 0.09
n 3.44 0.14
+
while for n-one obtains
B 2.0 0.1
n 1.5 0.2
pC
--..-..... . -+--+--+-
2
M (GeV/c 2)
Fig. 2: Plots of <p > versus mass for TI+ TI- and protons incident
on carbon. T
442 M.CHEN
I I I I I I
a
+
J -
1.0 f-
For p induced : ---T--- .J..
r
- - --.---+---+- -
p_w
0.5 - -
++
/ V
,II' I'I'
/
--+--+
-*----T--- 1
T
-
p_w
0.5 - -
I I I I I
0 2 3
M (GeV/c 2 )
Hodoscope
IEuIEuIE~ EulEull
I ~ t I
1 12 l T4 T5 T6
Spark chombt>r
Torgl't modutes
sc:int illotors
Beam
SI
Fig . 4
N N
V V
~ :><II
I.!) I.!)
N <'j
d 0
VI
$
c 1:
<II
<II
> >
<II <II
Q. ::.:::
tei Tt Tt' I::
I
p
0.1
n
0 .21 0.50 1 2 3 5 0.21 050 1 2 3 5
Fig . 5
446 M.CHEN
Since the ground state charmed particles are all stable against
strong decays and therefore are expected to have relatively long
lifetimes ( T <10-13 sec ), the most direct means of detecting them
is to observe a narrow peak in the effective mass spectrum of
several hadrons.
C 5
_ ~ ~\\\\Y
Beam I
/1
Target
Plan view of the detector
b 2m
Mo' HI ' M2 - dipole mngncts
Ao ' A, B, C - 8000 proportional wir chambers
a , h - 8x8 hodoscop s
S - 3 banks of pb-glass shower counters
CB, Co Ce - gas Ccrenkov count rs
Beam
~
Target
o 2m
Fig. 6
448 M.CHEN
, "
Sens~t~v~ty
(cm)
2
Ah x
7T
+ -
K
2.25 GeV
1 x 10- 33
3.1 GeV
4 x 10- 35
3.7 GeV
1 x 10-35
+ -
KIT 4 x 10-33 8 x 10- 35 4 x 10-35
p P
- - 4 x 10-34 2 x 10-35
+ -
KK 1 x 10-33 5 x 10- 35 1 x 10-35
+ -
7T7T 8 x 10-33 5 x 10-34 3 x 10- 35
.
(.)
......
>Q>
~
\0
C\i
......
~
C
Q>
~
0
eo
40
Fig . 7
450 M.CHEN
+
d 30 x 10- 30 for K IT
E .B <
6.
+
dp3 2.1 x 10- 30 for K IT
P .73 GeV/c
T
x .32
earlier.
+ -
(b) In the Wh h mode
+ -
As we discussed in (a) charm was not observed in the h h mode.
There are however two possible explanations that charmed particles
may be produced abundantly without giving significant signal in the
+ -
h h mode:
NEW PARTICLE PRODUCTION IN HADRONIC INTERACTIONS 451
l Lu
-
p + P +D+ D + x
or Kp if Baryon
KW or A ~\i
2
cm
452 M.CHEN
P ll
PJO
P5
--
--r-
--
95
1.5
~ Steel
III Hevimet I$cint
eTR
fIT] Beryllium
1.0
0 .5 O eH ,
~ pwe
o 25 30
Meters
do -37 2
B/ 3.4 0.3 x 10 cm /GeV/nucleon
dy
y = 0
do -37 2
B. dy I = 1.86 0.1 x 10 cm /GeV/nucleon
y o
NEW PARTICLE PRODUCTION IN HADRONIC INTERACTIONS 455
l- I I I
I- -
+
I-
++ +++ +
I-
++ +
I-
++++++ +
I- -
++++
-
l-
t
o +H
t
II
>-
71- -
-
Nbl~
E
I-
I- -
"0"0 I-
tttt
I- -
I- -
I- -
f-
l-
tttt
I I I I
9 10 11
mlJ.lJ. (GeV)
Fig. 9. Dimuon mass spectrum measured with 400 GeV proton beam
at FNAL.
456 M.CHEN
-37 2
B. dO, = 0.74 0.07 x 10 cm /GeV/nucleon
dy
y o
2
with X 18.7 for 18 degrees of freedom.
The Physics
(1) p + p +
Yv + X
L j/ \1-
(2)
(3) Looking for more J-like new particles, ~.e. vector particles
with new quantum numbers other than charm. We believe there are
more such particles in the mass range 10 to 50 GeV region. Vector
particles, similar to photons, decay into lepton pairs.
~-- .....
-..
p + P ZO + x -..
L,u fl-
~E Ulal ~
I'OKE YOKE
magneti zed , ,
qz,ozzTL1~"'-".~""'&
P
LUZ7//7Z7ZZZ7771lL21
.~~~~
., " ,,
", ,
----:
I s:
" I
,, I (")
I '- :I:
m
Z
Fig . 10 . Dimuon detect or at the ISR . niv . 441.60
NEW PARTICLE PRODUCTION IN HADRONIC INTERACTIONS 459
10000
VS =62
.. VS =52
1000 -
u
----
:>
0.> NU
'-" 100
'-
0.> ----
:>
0.>
0- '-"
....,VI '-
0.>
c: 0-
0.>
>
UJ ....,VI
10
"""-
C
0.>
>
LLJ
"-
""
1
6 8 10 12 14 16 18
"20 1
20 25 30 35
m"" (GeV/c 2) mzo (GeV /c 2 )
a b
Fig. 11. Expected dimuon events from Yv (a) and from ZO (b).
NEW PARTICLE PRODUCTION IN HADRONIC INTERACTIONS 461
oDD.
RUN ND . 715 EVENT 176
" ASS-16.S39 0.165 "fAP. 2.695 xr 0.575
p\- to . 'll! Xl- \.579 P2 --IQ.617 X2- 5 . 971,1
lOO.
MUON
......
100.
1.4
o. I IIU e
,100 .
2DO.
)00 .
REFERENCES
1) J.J. Aubert et a1. Phys. Rev. Letters 33, 1404 (1974). See also
e+e- Physics, same proceedings, by H.F. Schopper.
2) 11.Chen, Lecture at International Summer Institute in Cargese
on "Weak and Electromagnetic interactions at High Energies"
published by Plenum Press, 1976, Editor N. Levy.
3) B.C. Brown et a1., FNAL preprint (1977)
K.J. Anderson et a1., Phys. Rev. Lett. 36, 237 (1976)
B. Knapp et a1., Phys. Rev. Lett. 34, 1044 (1975)
4) C. Kourkoume1is, CERN 77-06 (1977)
5) S. Ellis et a1., FNAL preprint 76/29 - THY
T.K. Gaisser et a1., Phys. Rev. D13, 171 (1976)
6) 11. Binkley et a1., Phys. Rev. Letters 37, 578, (1976)
7) J.G. Branson et a1., Phys. Rev. Letters 38, 580, (1977)
P. Piroue, B.N.L. preprint (1976)
8) V.A. Matveev et a1., Lett. a1 Nuovo Cimento 7, 719 (1973)
S.J. Brodsky et a1., Phys. Rev. Letters 31, 1153 (1975)
9) J.G. Branson et a1., Phys. Rev. Lett. 38, 1334 (1977)
J.J. Aubert et a1., Phys. Rev. Lett. ~,416 (1975)
10) G. Fanar, Phys. Lett. 56 B, 185 (1975)
11) S.D. Dre11 et a1., Phys. Rev. Letters 25, 316 (1970) also
Ann Phys. (N.Y.) 66, 578 (1971)
12) A. Donnachie et a1., Nuc1. Phys. Bl12, 233, (1976)
H. Fritzsch, Ca1t 68-582, (1977)
13) S.D. Ellis et a1., Phys. Rev. Letters 36, 1263, (1976)
14) U.J. Corden et al., to be published
15) F. Bourgeois et a1., CERN/DD/DH/70-1
16) J.J. Aubert et a1., Phys. Rev. Lett. 35, 416, (1975)
17) D. Bintinger et a1., Phys. Rev. Letters 37, 732 (1976)
18) U.G. A1brow et a1., Nuc1. Phys. Bl14, 365, (1976)
A more recent upper limit from ISR (R606) sets BO' < 9}JP.
see ISR workshop/2-8/1977, Editor U. Jacob
19) M. Chen. Proceedings of the Vanderbilt Conference on New Results
in High Energy Physics,_ Editor R.S. Panvini (A.loP., N.Y., 1976)
20) S.W. Herb et a1., Phys. Rev. Lett. 39, 252 (1977)
DIS C U S S ION
DISCUSSIon
SANDA:
CHEN:
SMITH:
JUNG:
I was on the n.
experiment. On the number u-{J<+)/rr(lt)
= 0.85 Z. 0.5, the error includes the statistical error
and a systematic error of 10%, which we think to be
463
464 DISCUSSION
CHEN:
MIETTINEN:
JUNG:
CHEN:
BACE:
CHEN:
BACE:
JUNG:
The results are very sensitive to the quark
distributions used.
CHEN:
The data points are between the two hypotheses,
with and without color.
OLSEN:
In the graph you showed for the cross-section for
pp -+ JX, the cross-section at 40 GeV was 4 or 5 times
the cross-section at 30 GeV. This corresponds to a
change in CM energy of only 1.5 GeV out of 8. In going
from pp to trp, to pp the amount of available CM energy
changes by a comparable amount. Does it make any sense
to' compare the J production cross-section for differ-
ent incident particles at the same laboratory momentum,
when the cross-section is changing so rapidly?
CHEN:
I agree there is some problem. I think that at
least the cross-sections for pp ~ JX should be plotted
vs rs -2 'l"4J> to account for the limited available
energy, as was already indicated on the excitation curve
I presented.
PREPARATA:
CHEN:
In the charm quark or-Bluon fusion model I would
guess it should also be ~ - 2Np for the p case since
not all of the quarks are annihilated. Therefore the
ratio and the conclusion that it cannot be purely done
to these two processes remain to be valid. In the quark
antiquark annihilation model of Drell-Yan, it is not
clear which variable should be used. In the very extreme
466 DISCUSSION
MINGUZZI-RANZI:
CHEN:
WHITE:
CHEN:
JUNG:
WHITE:
CHEN:
JENNI:
CHEN:
CHEN:
MIETTINEN:
May I comment on this question of the mean trans-
verse momentum of dilepton pairs vs their mass. Figurei
shows new preliminary Fermilab data presented by Leon
Lederman at the Budapest Conference, two weeks ago. As
you can see, <ak is increasing with increasing ')'K.~..k
but it levels otf at a value of <&:>~ 1 GeV Ie. above "about
""~I' '" 4 Ge VIe. 2.
This is a very important new result. If we think
that the pairs are produced in an incoherent interaction
among the constituents of the colliding hadrons (such as
in the Drell-Yan process) then the transverse momentum
distribution of the pair reflects the transverse moment-
um distribution of the constituent in the hadrons. New
data on large p~ hadron-hadron collisions from the CERN
ISR also provides evidence that the <PJ.> of the partons
is much larger than the typical transverse momentum of
secondary hadrons, which is about 350 MeV/~.
MINGUZZI-RANZI:
I do not understand why J production via an inter-
mediate virtual photon after ~1 annihilation does not
contribute substantially to the cross section.
CHEN:
The fact that the ratio of the J signal to the di-
lepton continuum is much bigger in hadronic production
than that in e+e- annihilation shows that the direct
coupling of,; to the J dominates over the single photon
intermediate state.
JENNI:
I believe that a more sensitive way to look for
charm production in hadron interactions is looking for
semileptonic decays by looking for ~~ ':f= coincidences
from cC: ; the branching ratio l> ~ K+ l y <'OJ 20% while
the branching ratio for D -+ Klr is only ,.., 2%. This
DISCUSSION 469
1.5 -
+
>CII
1.0 f-
t t+++ 1
T
f
(.!)
::1.
::1.
'"
~
**
x
........
Pilcher et al VS = 16
0.5 -
This experiment V5 = 27
I I I I I I
0 2 4 6 8 10 12 14
m l1lL (GeV)
SMITH:
CHEN:
A.J.S. Smith
1. INTRODUCTION
In the last year much progress has been made i'n studying lep-
ton pairs. Not only is there an exciting new structure at 9.5
GeV pair mass 4 ), but also much better data on continuum pairs.
Here one observes scaling behaviour, as predicted 5 ) by Drell and
Yan, and also a striking similarity in the properties of the ~n
f. P - 'J"I'" -t s~s + )(
L,.. "lti'1t+n.-7t-nO
aN
dc.o~e~ ... "2
"
where e* is defined in Fig. la. In terms of the transverse
momentum PT ~ ~M . e*
S1n (neglect masses of final-state par-
ticles), we get
(a)
L
~
..
Beam Direction
/
..
/
II
Missing Momentum
In Center of Mass
Of Particle M
I
(b) I
J
3 J
I
J
I-
.'.
I
0- I
..
~ 2 I'
Z
"0 ~
,
....... , ".- .- ' I '
...' .".'
J '.
I
I
I
" ...... I
J ''
...... J
........ -....
'
...
M/4 M/2
Observed
Muon
Yield
Direct
Yield
1119 (Uranium) up
Fig. 2 Method of ex tracting direct muon rates from measurements
with absorbers of various densities .
o 10 20
I I I
Tor9" scolt 1ft mel"e
Even after correction has been made for the above backgrounds,
there remain real prompt, direct leptons from dull sources; for
+ -
example, the decays of Po, wo, <Po -+ Q, Q, In v~ew of the advanced
+ -
age of the J/1jJ particle, the decay J/1jJ -+ Q, Q, must now also be
classed as a pollutant. Measuring lepton pairs yields the cross-
sections for vector-meson production -- interesting physics in it-
self. However, the copious production of prompt lepton pairs is a
limiting factor in searches for new phenomena with single leptons.
6 7)
The first definitive experiments came in 1974 from FNAL' ,
8) 9) 0
the ISR ,and Serpukhov . All groups looked near 90 in the
centre-of-mass, or x F ~ 0 (the so-called Feynman x is defined as
x F = 2 p~//S, where P~ is the c.m. longitudinal momentum, and IS ~s
the available c.m. energy). Here the background rates are lowest,
and most new phenomena should occur. Experimentally it is often
easier to measure the relative yield Q,/n rather than the more fun-
damental quantity E d 3 0/d 3 p. For high PT the Q,/n ratio is also a
natural scale factor, in that both Q, and n come from hard scatters.
Also, this ratio facilitates comparison of experiments having dif-
ferent target nuclei. The basic experimental result is simply:
-4
Q,/n ~ 10 for both e and ~ for PT ~ 1.3 GeV/c. In the forward
direction, Q,/n is not as useful a quantity as E d 3 0/d 3 p because
large numbers of pions are produced peripherally in this region;
also the lack of comprehensive n cross-section measurements makes
it difficult to extract E d 3 0/d 3 p from Q,/n.
200 GeV). They measured the muon yield for various secondary
momenta by inserting a hadron absorber between the target and the
first quadrupoles. Typical results are shown ~n Fig. 4, where
the invariant cross-section E dcr/d 3 p is shown as a function of PT'
One sees that over seven decades E dcr/d 3 p for ~+ production is al-
_4 +
most exactly 10 times the n cross-section in the same region of
phase space.
versus PT' For PT > 1 GeV/c one sees similar behaviour to the
Fermilab data. For lower PT' on the other hand, ~/n increases
dramatically, according to the ACHMRN group. One should note
o 0
that their measurements are at 30 c.m. rather than 90 , where the
CCRS and Fermilab measurements were made. As we shall see, such a
large increase at low PT is difficult to explain phenomenologically.
The dependence of ~/n upon the c.m. energy is is shown in Fig. 7
for PT > 1.3 GeV/c, and in Fig. 8 for lower PT: 0.4 ~ PT ~ 0.6
GeV/c. The data of Fig. 7 are reasonably consistent and seem to
indicate a threshold at is ~ 6 GeV. The situation for low PT
(Fig. 8) is not at all clear, however, and more experiments seem
to be needed to clarify things. I point out that the low x F ' low
PT region is certainly the most difficult for experiments: the
Dalitz-decays are copious, and muon experiments are almost out of
the question. Hence we shall devote most of our attention in in-
terpreting the prompt lepton data for PT > 1 GeV/c, where at least
the experiments are reasonably in agreement.
480 A. J. S. SM ITH
\
1\ 300 GeV
p+nucleon-,u+
"\
~ 10- 34
0
"-
> Q)
(!)
I
I
\
N
E
u
10- 36
,..,: .
\
'0
"-
b
'0
W
1,
10- 39
0 2 3 4 5 6
Transverse Momentum ( GeV/c)
['8
3.0
!
2.0
<;t
0
x
1.5
--
t:::
~ 1.0
0.5
0
1.0 4.0 5.0 6.0
Transverse Momentum (GeV/c)
P.L (GeV/c)
Fig. 6 The electron to p~on ratio at ISR energies, vs PT
Note that the points of Ref. 11 were measured at
8* = 300, whereas those of Ref. 12 and Fig. 5 were
measured at 8* = 90.
482 A. J. S. SM ITH
o et (CCRS)
1.3< P1 < 3.0 GeV/c
,.,.-(CP)
60 0 < 8*< 100 0
e ,.,.- (CP), et (CF)
2.0 v ,.,.~ (CHPW)
,.,.+ (Serpukhov)
e+ (P-SB)
1.0
2.0r- -
v
o
x
1.0 -
Fig. 8 Dependence of 9-/rr upon /S, for measurements with 0.4 < PT
< 0.6 GeV/c.
LEPTON PRODUCTION IN HADRON-HADRON COLLISIONS 483
~
(!)
......
N
E
u
..,a.
~
b
"0
W
-39
40 O'--'--2'--3'--4L-J5L-J6L-J
Transverse Momentum
( GeV/c)
10- 3
o Columbia -FNAL
Chicago - Princeton
Ft~~j#1
1(j4
10- 5
I::
"- 10- 4
cy Total
---
10- 5
PT (GeV/c)
~ CCRS
?9 0 CHORMN -30
PT (GeV/c)
gluon
q r
( b)
Finally, there could be some very dull sources not yet included
in the estimates of Cronin or Bourquin-Gaillard, such as the
o +- 0 +-0
Dalitz decays n + ~ ~ y or w + ~ ~ ~ , which have not been sub-
tracted from the observed muon yields.
Accompanied
. ::: 1.1
All smgle j.1
In making the large correction for their acceptance, they have as-
sumed that the j.1-pairs decay isotropically in the dimuon centre-of-
LEPTON PRODUCTION IN HADRON-HADRON COLLISIONS 489
> 0.2, PT < 0.6 GeV/c) comes from pairs. The average mass of the
pairs is <M> ~ 900 MeV, and the ~-pair yield from p, , w is
times too small to account for their signal. The resolution of
this experiment lS too coarse to see the individual vector-meson
contributions.
Lead Walls
r P Hodoscope
j
Spark Chambers
Target
Chicago Cyclotron
Magnet Iron Hadron
Absorber
Fig. 13 Spectrometer of the Chicago-Princeton II group.
150 GeV
P"Se--fLfL+X CP-lI
~ XF >0.15
o
~ 10
u
::l
Z
:;;
Q)
t!>
"-
I
~
~
~
b
"U .1 ("7-fLfL y)+ (w-fLfLTT")
(0""7 = O"w =O"p)
~
'()
492 A. J. S. SM ITH
Using this fitted form, we then summed over all pairs to obtain
the inclusive single-w yield. The results, as shown in Fig. 15,
are: pairs from vector mesons alone cannot account for the ob-
served win ratio of ~ 10- 4 , consistent with Cronin's calculation;
the inclusion of the previously undiscovered low-mass continuum
changes the picture - the single-w yield at x F = 0 can be explained,
for 1.0 < PT < 2.5 GeV/c, by the measured spectrum of w-pairs. A
similar conclusion applies in the forward direction, as seen in
Fig. 16.
Since last year, many checks have been made on the data, to
be sure that the non-resonant continuum is real, and not due to
secondary W production or uncorrelated background tracks. Measure-
ments made with different target thicknesses show identical mass
spectra, indicating no secondary sources. Also, tracks downstream
of the hadron absorber from one event were matched with upstream
tracks of a different event, to estimate the background from un-
correlated tracks. Finally, careful Monte Carlo estimates were
made of background from the Bethe-Heitler production of pairs by
secondary photons. Less than 3% background from any source was
found, so we are sure this low-mass continuum is established. It
cannot be explained by conventional sources such as nO ~ WWy or
WO ~ nOww, which fail by an order of magnitude to account for it.
The solid line in Fig. 14 shows the shape of the Dalitz spectrum --
it is clear something else is needed!
. 23) .
A recent exper1ment uS1ng the SLAC streamer chamber has 1n
fact confirmed the existence of this low-mass continuum. They have
+
looked for inclusive w-pair production in 16 GeV/c n-p collisions,
and have measured the spectrum (preliminary) shown 1n Fig. 17. A
very similar shape to that measured by CP II 1S seen. The authors
also conclude that an unconventional source must be responsible.
LEPTON PRODUCTION IN HADRON-HADRON COLLISIONS 493
4tJ !
l-
----
..Cl
e
~1.5
a.
'"-a
t
~
1.0
----b
'"-a
W 0.5
~!~ ____ l_~
0.1 0
2 3 0 I 2 3
PT (GeVlc) PT (GeV/c)
1
1.4
\
\
,
,, ""0 1.2
\
\
,,
.l
c
f
-u.10 3 , ,x 1.0
><
-a "- I-
"-
----b Vector Mesons>'" ~ 0.8
~
-a
Only \
,
\
0.6
\
\
\
, 0.4
\
\ 0.2
\
---- - -- ---- -- - ------
10
0.1 0.2 0.3 0.4 0.5 0.6 0.1 0.2 0.3 0.4 0.5 0.6
XF XF
Fig. 16 Contributions of V pairs to the yield of single
~'sin the forward direction.
494 A. J. S. SM ITH
1T + P-fL+fL- + anything
16 GeV/c. XF > 0.3
>Q)
::0:
o 10
~
-en
c:
Q)
>
w
(."_y,.,.+,.,.-)+(w_.,,.o,.,.+,.,.-)
CT." = 1/2 CTW
loml"-__________
100m,
Fig. 18 Multihole spectrometer of the CP I group at FNAL. The holes
are filled with large liquid scintillators. The authors
trigger on a high-PI particle in the precision spectrometer,
and obtain a rough measurement of the energy and angle of any
muon penetrating the earth on the opposite side of the beam.
LEPTON PRODUCTION IN HADRON-HADRON COLLISIONS 495
1) West of the Iron Curtain, for beam energies above that of the
Brookhaven AGS (IS ~ 8 GeV) , and for PT ~ 1 GeV/c, single lep-
tons come predominantly from pairs of electromagnetic origin.
The same conclusion holds for the forward direction.
.../e+e-
-29
w
-30
-34
JI'/I
N
E -32
u
.-c: -33
E V/-L+/-L-
"C
I
"- -34 I
b I
"C
I
-; -35 I
.2 I
I
I
-36 I
I
I
I
-37 I
I
I
I
- 38 I
I
I
-39
.4 .2
en
~
Z 40 ~
W
>
W
CPII
225 GeV Ie
p+C-f-Lf-L+X
(Preliminary)
t
This I.pl .
Hom I' 01
:;-: 0 Kluber9 el
~ ~
\fW
f
~ ~
10000
~"
j IOOOO} .00.
b)
~
u
::J
c=
.......
~ 10000
N
U
.......
>C1I
<.!)
1000
.......
..0
c=
100
10
0)
10- 30
,,
II) ,
~
\',
.!! '. , J/'"
u
~
Z to- 31
........
<II
U
........
>III
(!)
........
N
E
10- 32 i "~Jt
u
3000
NUMBER OF
EVENTS WITH
2.5 <M<5GeV
1000 1250A
500
300
100
50
Ii' 3
-)8
10
6 10 12 1"1 /6
}foss G.... V
Fig. 22 Results of Herb et al. (Ref. 4) for d 2 a/dH dyly=o
in 400 GeV p-Nucleus collisions. The inset shows
earlier results for lower pair masses.
LEPTON PRODUCTION IN HADRON-HADRON COLLISIONS 505
--- ---------7
1.2
1.0
.8
~
.6 1+ 0.02 cos 2 (J "\
'c
:::> .4 \
IT1
....
>- .."
,
EFFICIENCY-----,\
c .2
.... \ ~
..... ()
:0
.... IT1
z
5 J/'i' MASS, 17":!:,p on e8 Sn ()
1.2 -<
<J:)
en 1.0 1.0
0
u .8
"0
....... .6
b
"0 .4 ----- "
.....
.2 "
EFFICIENCY~",
0 "
e
0 .2 .4 .6 .8 1.0
cos
(f)
~ 103
W
....J
U
~
Z
......
(f)
Z 102
fl::
CD
0
z
z
> 103
~
(!)
lL.
X
"0
......
b
"0 102
W
XF
Fig. 24 b) The M ~ 1 GeV region.
LEPTON PRODUCTION IN HADRON - HADRON COLLISIONS 509
I I I
E dO" _ (1-X )N
dX F F
pC
7 o rr+C
6 rr - C
5
N 4
3 t
f ! J
J I t
2 t ( i
~ ~
~
~
I
1 2 3
M (GeV/c 2 )
The absorption of the J/~ itself has been taken into account (it
is very small). Within large errors, the higher-mass data of CP I
are consistent with a transparent nucleus. As I understand this,
low-mass hadrons are produced by highly-probable soft scatterings.
High-mass objects, on the other hand, come from hard scatterings.
The fact that the beam particle has previously undergone an ordi-
nary low PT interaction does not remove it from the "beam", and it
(or its fragments) can still undergo a hard scatter further into
the nucleus. Although not shown in the figure, a similar picture
has been observed by CP II for pion-induced pairs.
LEPTON PRODUCTION IN HADRON-HADRON COLLISIONS 511
"-
(/)
z
a::
~
CD
o
Z
.~
Z 10 2
Q.-i
'0
"-
b
'0
10
Q.-i
"-
w
1.0
1.5 o BRANSON el 01
Do HERB el 01
0
i
'"> ( ~ ~
Q)
(!)
1.0 ~
d!
A-
~
v I
0.5
2 4 6 8 1O
M (GeV/c 2 )
I I I
1.0
0.5 -
Branson et.ai.
x Binkley et.a!.
o Hom et.a!.
DAntreasyan et.a!.
o 2 3 4 5 6 789
MfLfL (GeV/c 2 )
1
p-w
-+-t -+- ++ t J/Ij!
AI
LL
X
~
'01
......
b
o
~
0::
~5~--~--J---~2~--L-~3~--L---
M (GeV/c 2 )
100
10
'2.
-ft~') = C e~p (-14-.7 M )
s .Ji
1
200
516 A. J. S. SMITH
o Ref. 34
Ref. 4
N
> Q) ~0-33
t9
I
N
E
u
>.
"C
~
"C
.........
b 10-35
N
"C
ro
~
~-~L-__~__~____~__~____~__~____~
o 0 .2 0.4 0 .6
M/.IF
I, 0
x
Ref 3
Ref. 36
0 Ref. 26
1032
tIll ~ "
Hom et al (Ref. 34 )
Antipov et al (Ref. 37)
Nagy et al (Ref. 37)
(; CCRS (Ref. 42)
N ~0-33
E
u ,, i J/t
0
,,
"0
>.
46 34 "1,
f
If,,,
........
b
"0
ro
~O-35
,,
\ ,
~0-36 \
0 0.6
n + A + J/~ + D + D + anything
One can search for the latter process by measuring the rate of
+ -
three-muon events, two muons coming from the decay J/~ + ~ ~ , the
third from the semi-leptonic decay DO + K~v, which has a 10% branch-
ing ratio. The background is, of course, muons from decays of
ordinary hadrons produced along with the J/~. As described earlier
in this lecture one tries to absorb as many hadrons as possible be-
fore they decay. The CP II group reached a limi t 39)
o(J + charm)
O) < 0.01 w~th 90% conf~dence .
o (J + anyt h ~ng
Acknowledgements
REFERENCES
5) S.D. Dre11 and T.M. Yan, Phys. Rev. Letters 25 (1970) 316.
10) These figures are taken from Ref. 1), where direct references
to the individual data can also be found.
14) M. Bourquin and J.M. Gaillard, Phys. Letters 59B (1975) 191.
15) G. Farrar and S.C. Frautschi, Phys. Rev. Letters 36 (1976) 1017.
36) All data on this plot have been referenced previously except
that of J.H. Cobb et a1., Phys. Letters 72B (1977) 273.
DISCUSSION No.1
PAAR:
SMITH
PAAR:
To get back to the Drell-Yan calculation, I believe
that color was not included in that calculation.
ANTREASYAN:
MIETTINEN:
WHITE:
SMITH:
CHEN:
HINCHLIFFE:
SMITH:
HINCHLIFFE:
SHITH:
CHEN:
HIETTINEN:
JUNG:
SMITH:
6
We had 10 part/sec: that leads to 3-4 10 part/sec
in the chambers. The beam spot was about 2.Sx3 cm 2 and
the electronic dead time was about 500 ns, so from this
side things turned out well (after a while). A problem
we had foreseen was possible back-scattering from the
absorbers, but it turned out that there was no serious
back-scattering. So, once we understood this, it caused
no problem. About half of our data include the inforrr~
tion from these chambers, and that information is really
only important for low-mass resolution.
WHITE:
I have a comment on your graph showing the s-depend-
ence of "low-pJ." lepton production, where "low-p",," is
supposed to mean 0.4< p",,< 0.6 GeV/c. In CCRS, our lowest
p~ bin starts at 0.6 GeV/c, so our point should not really
be included. If the lepton-to-pion ratio is increasing
below 0.6 GeV/c, it would clearly be wrong to do so.
SMITH:
I did not mean .4 to .6 GeV/c as a bin. I was only
indicating the approximate domain of PJ. from which these
data are taken.
CHEN:
The huge ~ peak in the dimuon mass spectra looks
very familiar to me. In a 1969 experiment on photoprod-
uction of e+e- in the f> ' ro region we found that the CO
peak is more than a factor of 2 greater than the r
whereas we know that ",-2: Yi2.= 9: 1. So here it's
not necessary to have m~re ~than p produced but we can
understand this as an interference effect.
You showed a scaling curve for the J and the ,.+p. con-
tinuum where the two sets of data were the same except
for a factor of 200. To me this implies that vector mes-
on (~. Col. C, :J ) production mechanism and the Drell-Yan
528 DISCUSSION
SMITH:
ZICHICHI:
PREPARATA:
I think it is clear that it cannot be a single re-
sonance. We can estimate the expected rate for r (1'... ~tl)
and we have this measurement of 1.2 GeV for the total
width. So its branching ratio would be so small
(10-+- l~s ) that it is unlikely that CFS would have seen
it. On the other hand if it is anything like the r it
is very reasonable to expect that there will be a total
of 2-3 states with small separation.
MIETTINEN:
SMITH:
WU:
Can you make a consistent picture of the A-depend-
ence of the dimuon production spectra and the A-depend-
ence of inclusive single muon production as measured by
CP-l?
SMITH:
ANTREASYAN:
SMITH:
WU:
I would like to have a clear picture of the J prod-
uction mechanism. Can you explain the relative shapes
of d~(?f'~:n<.)/dxl=' and dCT\1t"I>~:rX)/dxF ? You
showed that at small x these cross sections are approx-
imately the same but at large x l=' the "It. cross section
becomes larger.
SMITH:
Xb-XT=X F
2
')tbXT =.!!...
s
'::1 9/400
for our experiment, where Xb is the x F of the beam par-
ticle parton and x T is the x F of the target parton.
530 DISCUSSION
JUNG:
SMITH:
MIETTINEN:
PREPARATA:
MIETTINEN:
ZICHICHI:
PREPARATA:
BACE:
HINCHLIFFE:
If you believe in scale invariance you would expect
that <~L~ would be directly proportional to Q2 - the
only other scale in this picture. So you would expect
that <?L> would continue to grow with mass. However, as
Feynman has shown in his book, this also can depend on
R = a L /o-T and you get the relation <?J.'2.) ~ ~z'R.
When you calculate R in an asymptoticallY free gauge
theory, R(Q2) falls rapidly as Q rises so that
< 1>l"> ac Q2 'R.(Qt) can start to level off.
PEREZ-Y-JORBA:
It seems that one can fit the cross section
M? d~/dmd::J I ~=o for the production of lepton pairs
by very different functions of M2/S J one being exponent-
ial in M'ls the other one exponential in ""/Ji. Could
you comment on this? I also would like to know whether
there are theoretical predictions on this functional
form.
SMITH:
First of all if you look at the Chicago-Princeton
data I think you will agree that the error bars are
probably large enough to accommodate a fit of either
type. The new and more precise measurements of CFS can
probably not fit the form ex\, l-a. MVS). At any rate
the agreement between the data and the form ex~ l-b M/.JS)
is very striking to me. There are probably theoretical
predictions derived from parton distributions measured
in deep inelastic lepton-hadron scattering for the
distribution, but I think at this time the fits are
purely empirical.
NARRm-l RESONANCES IN BB REACTIONS
S. Nilsson
1. INTRODUCTION
533
534 S. NILSSON
are new results. Most results have been obtained in the last few
months. Being new they need confirmation. The statistical signi-
ficance is, however, in most cases sufficient. Taken together, the
large amount of new data indicate that something is happening. It
could be that the long expected two quark-two antiquark baryoniuffi
states at last have been discovered. We will start by explaining
what baryonium Llight be.
Let us begin with the quark structure of matter. The quark
content of mesons, strange mesons and baryons is shown in fig. 2.
While unified theory works well for the meson strings, baryons
resist a unified theoretical description. G.C. Rossi and
G. Veneziano have made progress in the possible description of
baryon dynamics in dual and gauge theories (1). One result is that
the baryon resembles a Y-shaped string. This is illustrated in
fig. 1. The graph does not represent a number or a process, it
illustrates the structure of a mathenlatical operator, the coupling
of colour. The baryon is not A- shaped or a string with three
equally interspaced quarks. It is obvious in fig. 2, that in add-
ition to the three quarks, the baryon is characterized by a junction,
where only the gluons couple. For the string junction there is a
QCD interpretation in terms of certain colour indices flowing
through the Feynman graphs (1). So in addition to the three quark
flavour lines we will draw a fourth dotted line to remember this
fact. We are not interested in all possible theoretical views of
junctions in this paper, but note that it is a useful concept for
bookkeeping, when one is dealing with baryons.
We illustrate the lowest order diagrams contributing to anti-
baryon-baryon scattering in fig. 3. The left hand side diagrams
illustrate from top to bottom the contributions from the Pomeron,
one meson (qq), two meson (2q 2q) and three meson (3q 3q) exchange.
Duality was always in trouble with BB scattering. J.L. Rosner
NARROW RESONANCES IN SB REACTIONS 535
Mass
GeVlc'
~r------------------------------'
decoy modu
3
-
U
--
T
2 _PI>
5
- - - - - - - - - - - -pp -Ihr.-~Id- - -
L
u
I u
u a
d
I
Mesons
u
Strange mesons
p n
Baryons
noted the difficulties already in 1968 (2). The one meson exchange
in the t-channel was dual to a four quark structure (Zq Zq) in the
s-channel. This four quark structure has been with us for ten years
and now we are reaching the heart of the matter. What ~s this four
quark structure? Is it a physical state? Does it exist? Does it
decay? How?
In view of what was said before, we will be careful and also
draw the flov1 of the colour structure, the junction line. This ~s
done in the figure. The dotted line says "I remember I was a
baryon". By s- and t-channel symmetry we now discover, that we
have a whole new set of true dual diagrams (right hand side of
fig. 3) which also contribute. These now represent annihilation ~n
tion, the lower index for the number of quarks. This object is
being exchanged in the t-channel in the dual diagram giving an
intermediate s-channel (qq) mesonic annihilation state. Going to
the other two scattering diagram Zq Zq t-channel exchange correspond
J
to a state with qq and Z junctions in the s-channel, MZ ' and fin-
ally 3q 3q - exchange to M~ (no quarks in the s-channel but two
junctions). This is illustrated in fig. 4.
At last we have defined what we mean by "baryonium". It is the
J J J . .
M4 , HZ' MO objects illustrated in fig. 4. G.C. Rossi and G. Venenano
have estimated the trajectories for the M-particles and the inter-
cepts for t = 0 are also given in the figure.
An earlier model of the baryon was suggested by D.B. Lichtenberg
(3), where the baryon structure is an antidiquark (Q) - quark (q)
pair. The diquark Q = (qq) being two antiquarks. The Qq model of
the baryon (fig. 4, bottom) is a special case of the Y-shaped baryon
with one degree of freedom less. The quark-antidiquark system was
538 S. NILSSON
Scattering Annihilation
-------------
'II
dual
..
l[
---_/ '---
]
------------
[ ][
._-- ""' .. _--
- ---------
]
------------
f[ ) [
] [ ,---
I
tI
"- -
q q
juncllon Int.rc.pt
I
---------
--------- Mt :( 2q 2q) a 4 .. --12
\ no of quarks
q Cj
M~ =(q q)
T:
(antldiquarkl
Qq mod.1 of baryon
'fl,
-
----
'D,
- 'P, ---'P,
---'P,
10 0 ----'P,
----'P,
___ ~51
=1~~
_ _ _ _ 'S,
183
'So
200 216
---
420
MeV --r
Shapiro et 01 Dover ct 01
------------
1897
I
3
21!..~'-'C-------'-'-.-:-+-.TT"""~
2 - prong
6 - prong
-40 -20 20 40 60
Q MeV
+ + 2
f a: 1
p + qlo p'n' (Q,q 1m)
+ + 2 2
Q H - 2m ~ (p + q) ~
pn 4m III
M 1936 1
r 4 - 8
I 0
J 2
12 6"i ( ji"p)
D
e
0
b
<J -4
-8
- 12
9 ~
8
6 ,,;, (h)
..,
e
b~
<J - ,
Moss Mevlc 1
pp-iin
24
t
t f Ht +
o tot + t +
f 0+ ~++!
.D
E 16
b 0
0
00 0
0
8
1900 1920 1940 1970 M MeV/C'
3. HASSES 1940-2500
Ue will begin this section with a few remarks on off mass shell
annihilation. It turns out to be the ideal tool to prepare the
initial and final states for discoveries of baryonium matter. We
show examples of triggers p + TI+ TI- + P and p + TI- in fig. 9.
This ensures that the ~++ 1S exchanged in the t-channel. With the
colour flow as in the figure, a true mesonic state will be produced
at the lower vertex.
Now suppose instead, that we have the colour flow as in fig. 10.
NARROW RESONANCES IN BB REACTIONS 545
p
n
~]e:~~_
d ...;;...... d
d d
U: x-
p ~ ............. ) d
p ~]L!C:IT-
H i x+
j5 ~ ..ii ...... .. d
:-:)RC:: :-
d ............................... d
-+-
pfl -+ pp
was investigated. The results are shown in fig. 12. The events
have been selected by the requirement of 1175 < m(PF Tf -) < 1300 and
cosS GJ < 0, where 9 GJ is the Gottfried-Jackson angle for the proton
in the decay of U -+ pp. This reduces the background.
In addition to the 1935 peak the experiments find peaks at
11 = 2020 3 r 24 12
and M = 2204 5 r 16 + 20
- 16
The most remarkable thing is that the peaks are so narrow.
The cross sections for the two-body reactions
State 110
1935 10 5 7 5
2020 15 5 30 10
2204 20 5
NARROW RESONANCES IN BB REACTIONS 547
tt- _ ..
Pf
1<
13 GeV/c tt-d
P n
P or n p p X
tt- tt
30
N
~
>II
~ 20
-C
o
N
III
~
ILl
10
The cross sections have large errors but the absence of NO (1520)
MO (2204) might be significant and might have to do with the quark
contents of the particles. In the spirit of this talk we want to
point out the possibility of an intermediate state M~
p /':,.+ -
pp.
pp -+ - 1T+
'IT
JPCI G +
M 2190 10 r 100 20 3 1
-- +
U 2480 30 r 280 25 5 1
4. MASSES 2500 - 00
o 'D
~
1/1
SJ
;i.
-
c:
"C
b
"C
cos (e-) cos (e")
1>1 2.60:1:0.01
r < 0.018
pp -
+ NUX
LK37f
-QQ + MM
-
leo
.
~
GO
- 4~-
".;
~
..
<
2 40 9
~ 40
~
! "~
C>
.... 200
t
0
0
.
....
c
20
.g
0
c 160
D
e
e O L-~~~~~~~~
2~ 2..0 26-
'0 M lie: ...... . ) CCtVIC ' ,
.
D
120
E
~
z
80
40
0
10 20 Z.:i
o + - 0 + _
Fig. 14 Hass of K 1f- 1f+ 1f- for 12 GeV/c pp +K 31f 31f X
s S
reactions in BEBC.
250 .
>OJ
::t: 200 .
,...,
0
......
150 .
..0
E
-
0
u
100 .
C>
~
.D 50 .
E
:>
Z
~~~~1.~6~~~2~.~
2----~~--~~~----~~. 0
MtK ... n-n --'
24 X-_ ppn-
20
..
~
>
::I: 16
on
~
.
l!I 12
c
ell 8
0
2000 2200 2UO 2600 2800
M(NNn) MtV/C 2
.
~
X--_pTl-n
>
::I:
0
..
2
l
20
H -+ fI. N
s
at 2.80 and 3.05 Gev/c 2 , fig. 17 (29). The peaks are not narrow
I'V 100 HeV. These states could be produced by exchange of baryonium,
but for the moment we just note their possible existence.
A most remarkable result comes from another CERN Omega experi-
ment with the same trigger but using TI instead
m = 2.95 0.01
r-o
a BR"'" 1 ~ b
appears (30).
If one looks at the decay modes by plotting the pp mass, fig.
19, with the events defined as the peak minus the events in the
wings, properly normalised, there is indication of decays
12 G.V/c KP-~NN(Tt) c.n
c.n
L.~f ~
130 I lin
oco 0'"
N M
110~ lill
, , L 11
~
>
..
C>
TOTA L ~p,IT ' MASS SPECT~M aT 16 ~(t " <:&62 ......,,'
.,z
q .,.", .,. a
0
-
.l!/
c p.
> ::::
~ ! ..... -
....
'0
D
t
E
;,
z
1
PEAK-WINGS
4
10
5
~
C)
vo O~~I~~--hHH=--~-trL---
6
a:: _ ns7T+(4996 events)
~ 7/'-P-PPF7T -/ p s7TO(3341 events)
~
z
w
~ 10
~~--~~~--~--~I~~
2.0 2.4 2.8
MASS PPF' GeV
-+- 2.20 1f
-+- 2.60 1f
++
2.95-+-Ll p
5. CONCLUDING REUARKS
has been the question if the annihilation cross section equals the
difference between the pp and pp cross sections
CJ
ann
= LlCJ = CJ (pp) - CJ (pp)
-
0.8
0.4
biN
"O:g
0
-II::
'--
<J -0.4
-0.8
0 0.5 1.0 1.5
b (frn)
i, S=o J PC = 0++ 3p
0
0
1 1+- 1p
1
2 2++ 3p 3F
2 2
i 1 3S 3D
l- 0
1 1
1 0-+ is
0
1-+
2-+ 1D
2
3S 3D
2 1 1 1
3D
2
2
3D
3
3
NARROW RESONANCES IN BB REACTIONS 559
1
MJ 0. 4 - - + t
4 2
MJ 0. 2 lt
2 2
1 1
UJ 0.0 + - + -t
0 2 3
J
These trajectories are also shown in the figure. The H4 trajectory
and G.F. Chew's S = 2 trajectory coincide. The gan~ now is to put
physical particles on the trajectories. We have many candidates
at our disposal but further progress can only come when we know
i!
'"
'"
1111111
5
3
I. "I' GeV'
;0
3 q
M~ exchange
..
., 3 Ii
3 Jet slructure
"
leading I erm
2q
M~ exchange 2 jet structure
;0 q
M~ exchange I jel sir ucl ure
--- q
Fig. 22 Structure of the s-channel in annihilation.
NARROW RESONANCES IN BB REACTIONS 561
Table 3
Summary of measurements
Decay
Mass MeV Width MeV or reaction References
1455 420 17
1660 y-rays 216 9 10
1695 183 7
1795 < 8 pd 36
1897 ~ 25 pd 13
1932 2 94 14
1933 ~5 pd 13
1936 1 94 15
1939 3 <4 16
~ 1960 ~ 80 -> 517 37, 29
1975 1 <2 -+ odd 17 38
1986 1 ~8 - KK .. 38
~ 2000 100 20 -+ 517 37, 29
2007 5 105 10 -+ KKl7 39
2020 3 24 12 21
2150 6 116 15 22
2150 30 200 25 3--1+ 23
2187 3 56 8 40, 24
2193 2 98 7 24
2204 5 16 +20 21
-16
2310 30 210 25 4++0+ 23
2335 6 132 15 22
2350 60 190 60 4++0+ 41, 29
interference
2359 2 1.65 12 24
2363 2 171 10 40, 24
-- +
2480 30 280 25 5 1 23
2600 10 < 18 -+ K 317 26
2670 ~O 28
2800 -+ AN 29
2850 pd 42
2950 10 ~O 30
3050 - AN 29
3050 pd 42
562 S. NILSSON
ACKNOWLEDGEMENTS
My task was made very easy by having access to the excellent talk
by L. Montanet at the Boston Conference (29) for which I am very
grateful.
REFEHENCES
CHAIRMAN: S. Nilsson
DISCUSSION No.1
BIZZARRI:
567
568 DISCUSSION
BERLAD:
BIZZARRI:
CHEN:
PEREZ-Y-JORBA:
NILSSON:
ROLANDI:
NILSSON:
SANDA:
ANTREASYAN:
NILSSON:
UEMATSU:
DE LA TORRE:
BALDINI-CELLO:
NILSSON:
CELMASTER:
BALDINI:
w, CJ and +.
e+e- annihilation width of other vector mesons, the J~
If this resonance was a vector meson, we ~
would expect a width of about 1 keY.
572 DISCUSSION
MANDELLI:
Do any of the existing models explain the widths of
these states?
NILSSON:
No.
PREPARATA:
I am disturbed that these baryonium ideas have so
little quantitative basis. I wonder also where are the
excitations which correspond to the new degree of free-
dom which the junction represents.
CELMASTER:
You showed graphs of the reaction pp ~ i + X. You
have interpreted the three gamma bumps as a glueball,
2 quark state and a four quark state. Why do you adopt
this interpretation and not that, for example, of a glue-
ball and its excitations?
NILSSON:
Certainly one could consider other possibilities.
CELMASTER:
Why don't you see this glueball in J 1'1'-+'( + glueball?
NILSSON:
One needs to know the signature for such a process
in order to look for it.
FIELD:
I want to make a remark on the single gamma observa-
tions and the corresponding bound pp states. Having seen
the data, I feel that its statistical significance is in
doubt and that we should wait for better data before
taking these states seriously.
DISCUSSION 573
NILSSON:
JUNG:
The people who did the experiment have told me that
the only conclusion they can draw is to go and do the
experiment better, with a second NaI counter to reduce
background and systematic effects.
MIETTINEN:
I would like to make a comment. For ten years,
several theorists have emphasized that duality and dual
models don't make sense if states consisting of two
quarks and two antiquarks don't exist. Rosner, Shapiro
and collaborators, the Japanese people, and Dr. Lipkin
have long advocated searches for these states. The
reasons for the rapid progress being made now are the
new experimental facilities, such as theJl spectrometer,
and the renewed experimental interest in spectroscopy
generated by the theoretical and experimental successes
of the charmonium spectroscopy.
NILSSON:
DEO:
NILSSON:
There are no theoretical predictions for the widths.
DEO:
NILSSON:
The trajectories come from multiperipheral models
for these objects. For details I would refer you to the
original papers of Rossi and Veneziano.
BERLAD:
A follow-up question: we are told that in dual models
the slopes of trajectories are universal unless loop in-
tegrations are involved. Is there someone in the audience
who can ~ell us how the low values ~~ y~ and oC~ ~ ';'~
were derl.ved?
NILSSON:
I must again refer you to the original papers.
CHEN:
Is there any theoretical explanation that the 2.02
GeV particle is ~roduced strongly in association with
&1230) and the N (1520), while the 2.2 GeV is in associa-
tion with ~(1230) but not with the N~(1520)? Could the
latter be due to phase space?
NILSSON:
It might be due to the production mechanism.
DISCUSSION 575
ZICHICHI:
What is the pion energy in this reaction?
NILSSON:
ZICHICHI:
COMBER:
NILSSON:
There is no problem: the plots shown include only
exclusive NN channels.
COMBER:
What then is the meaning of the "('r~..,rTr" TtTf+)('Tf'~rrTl)'\
which follows the (pplr) on the diagram you showed?
NILSSON:
WILQUET:
I want only to inform everyone that the 5~ peak ob-
served in the mass spectrum at 2.6 GeV in the
576 DISCUSSION
LIPKIN:
NILSSON:
MANDELLI:
ZICHICHI:
MIETTINEN:
-
~~ ~
'j
:i
't
~f
~
~
1r T rr - ---J;> 'SO ---? 'Jt'+1J'
Now, in the string model, the baryon consists of three
quarks and glue-strings as follows:
NILSSON:
MIETTINEN:
A
Y
--;:>
CD
Veneziano and Rossi call these baryonium states M2 and
MO, respectively. Personally, I think that the theoret-
ical understanding of these low-lying "baryonium" states
is much less advanced than that of the qqqq states. The
M2 state looks to me very much as the one-loop correction
term to the normal meson wave-function:
.
DISCUSSION 579
,
, "
Meson= + .... + high spin
o ,
' ..
S(1935) ~ ~ -
""-
*)
N. Cabibbo
The facts about the parton model are, by now, standard know-
ledge.
581
582 N. CABIBBO
5 = (1)
The parton model starts from the consideration that the weak
interaction takes place on a distance r ~ 1/1Q2 which is small with
respect to the size of the proton, and on a time interval T ~ l/iV,
where V = (pq) is a variable which, in the lab. system, reduces to
the energy transfer from lepton to hadron. These conditions sug-
gest that we are really looking at the elementary constituents of
the proton on a time scale which is short with respect to the time
scale of the proton dynamics. According to the parton picture, the
interaction goes in three steps. The first contains all the time
up to the interaction: during this time, due to its interval dy-
namics, the proton slowly moves from one configuration of internal
constituents to the next with no forewarning of the coming inter-
action. Then comes the interaction which depends only on the in-
stantaneous configuration of internal constituents and the known
properties of weak (or electromagnetic) interactions. After this,
there is a long aftermath during which the elementary constituents
rearrange themselves into the final particles.
LX, b )
IT
(2)
4i (x) = (3)
In the last few years, our group has evolved a picture of the
parton distribution within a nucleon, based on the following idea:
we assume that the nucleon can, even at large momentum, be thought
of as made of three quarks, which carryall of the momentum. These
will be called "constituent" quarks. The nucleon will then be de-
scribed, as in the naive parton model, by a wave function that de-
. bl es (l.
pen d s on t h e space varl.a l. l..
x, PT' = I .,
2 3) 0 f t h e tree
h
quarks, as well as on spin and isospin indices. We will have:
(4)
L P~ = 0
(5)
(6)
These will satisfy three sum rules which express the fact that in
a proton there are two p quarks and one n quark, and they carry
all the energy:
(9)
PARTON DISTRIBUTIONS AND THEIR Q2 DEPENDENCE 585
(10)
(11)
(12)
586 N. CABIBBO
(near x 1)
(13)
Here, I will not present in any detail our ansatz for the
wave function of constituents in a proton, which is discussed in
Ref. (I), but will only recall that it contains two adjustable
parameters: S which fixes the x distribution and b, a parameter
which describes the mixing of configurations of different symmetry
and gives rise to different shapes for the Po(x) and no(x) distri-
butions. The b parameter is used to fit the x behaviour of
F2 (en)/F 2 (ep), especially near x = 1.
4. PARTONS IN CONSTITUENTS
9
a q
9
picture suggests the general form that the ab(x) will have at some
large Q2. We see from the above diagram that the constituent a
will contain one parton with the same quantum numbers. the "valence
quark". plus a number of gluons and qq pairs. the "sea". We will
then have, if b is a quark:
(14)
The first term is the "valence" contribution and the second is the
sea contribution. In this expression, we have assumed that the
sea distribution is the same for all qq. This assumption can
easily be released by specifying different distributions for non-
strange, strange, and charm quarks. The gluon distribution will
be described by a third function:
(15)
Although the gluon distribution does not enter the expressions for
deep inelastic phenomena. it plays an important role in determining
the Q2 dependence. It can also playa direct role in applications
of the parton model to hadronic processes.
=
PARTON DISTRIBUTIONS AND THEIR Q2 DEPENDENCE 589
QCD and QED differ in one important way: the first LS asymp-
totically free, while the second is not. This will have no effect
on the structure of the equations obeyed by parton distributions,
but only on the Q2 dependence of the effective coupling constant
a(Q2) in the two cases and, therefore, on the asymptotic behaviour
of the distributions.
The parton picture Ln QED has a long history dating from the
Fermi-Weizsacker-Williams approximation scheme. Applied to brems-
(16)
e-----jC"
2
-t ... (i-X)
(17)
and obtain y(x) by integration over PT' The result is found to de-
pend both on x and on p~ax ~ E, We will then write:
""o.x
r (x It:) = Jf 012.fT ( (x p)I =
(18)
= (1+ (i_)()2.)t
t = ~ (19)
/// .
etc .
.2
,et c.
--0(
21L
(20)
In order to determine P
,we apply (20) at lowest order in
ey
a by using our previous computation in Eq. (18) which gives, at
order a, the y distribution corresponding to an electron distribu-
tion:
(21)
(22)
(23)
(24)
(25)
J
~
(28)
where
(29)
has been chosen to obey Eq. (26) for any A. We then substitute
(28) and (29) in Eq. (23) and obtain a result which is finite as
A -+ 0:
J ~~ Pe~ tf)e(~,
~
t)
~
(31)
(a photon of energy E) we can compute directly e , at order a,
through the diagram
This gives
(32)
If we substitute Eqs. (31) and (32) in Eq. (23), and compare the
order a terms, we find:
(33)
.(1:)
(34)
PARTON DISTRIBUTIONS AND THEIR Q2 DEPENDENCE 597
1
(36)
it):
(37)
REFERENCES
BACE:
599
600 DISCUSSION
ZICHICHI:
CABIBBO:
BACE:
CABIBBO:
2
In QCD when Q ~oo you also tend to sea quarks only; you do
it very slowly:
This equation tells you e.g. that in QED you keep on adding new
DISCUSSION 601
positrons with lower energy made through pairs because the right
hand side of this equation is positive.
In QED, is also increasing with t so that the rate at which
you produce pairs will be increasing. In QCD the rate of quark pair
production slows down but never stops. This is the difference
between QED and QCD. If you take a typical SLAC x - distribution
at a certain tl what happens is that in QCD you tend to accumulate
at low x at a decreasing rate while in QED you would do so at an
increasing rate. In both cases, if t2 > tl the structure functions
change as indicated in the figure 1. The results obtained from
these integral equations with the running coupling constant are
equivalent to the results obtained previously using the equations
for the moments of the distributions.
o i "
BUDNY:
Could you show how the quark distributions pO(x) and nO(x)
are de rived ?
CABIBBO:
'f'abC 1 1
(x p~,
2 2
x P.a. '
33)
x PJ,
602 DISCUSSION
analogous
CABIBBO:
UEMATSU:
CABIBBO:
2 a
In the presence of a fixed point you have a (Q) - behaviour
where a might be a small number. So from experiment you cannot
distinguish between this and 10g(Q2) - behaviour. You can only put
limits on the power a. But our point of view is really different:
We believe in QCD and compute what happens; if we find it agrees
with experiments, we are happy, if not, we have to do something else.
UEMATSU:
2
If the parton distribution function depends on Q then the
2
fragmentation function is also expected to have Q - dependence.
What is your opinion on this point ?
CABIBBO:
TAO:
CABIBBO:
The justification comes from the results you get. They are the
same as in the usual analysis. For example in QED once you have the
Weizsacker Williams approximation it is a good thing because you can
compare it with exact calculations. If you ask a mathematical
question you cannot ask for an intuitive argument. I gave some
intuitive argument for these equations. Incidentally those can be
derived from Georgi's and Politzer's equations directly, because if
you solve the equations for the moments and then make an inverse
Mellin transformation, you will arrive at those equations after a
complicated and mathematically more convincing tour. This is act-
ually the way these equations for quark distributions were origin-
ally found by Parisi.
DISCUSSION 605
PREPARATA:
HINCHLIFFE:
CABIBBO:
This is a big problem which Q2 you should use, whether the mass
of the lepton pair or s or anything between. The Drell-Yan process
is the less well justified example among parton model applications
606 DISCUSSION
CARTER:
How did you get a pion wave function from that of a proton ?
o 0
Are u (x) and p (x) the same in both cases? What tests do you
have of this wave function ?
CABIBBO:
o 0
u and p will certainly be different. For the wave functions,
I best address you to the papers quoted in the lecture. Given a
wave function, I wish to recall how 1t leads to a K
'Ie
-f mass
difference formula:
,2. ,
*)
P. Renton
Oxford, England
1. INTRODUCTION
607
608 P.RENTON
2. NARROW-BAND BEAM
torget I
1
.. .!!'~-o--
1
-----
400 GeVlc
protons
t t
Oecoy Steel shield InO with
1
BEBe
tu nnel 00 ps conlo l nlno
solid state detectors
are directed into a decay tunnel, where the TI and K mesons in the
beam decay to produce neutrinos. About 3% of the TI mesons and 18%
of the K mesons decay in the 300 m long tunnel and the main decays
producing neutrinos are
+
-rr- -+ f.-+
E =
:J. [EM - PM cos ~J
~ PM (!- mt~/tm~)
1+ (PM/IMM)2f12
TOTAL CROSS SECTIONS OF NEUTRINOS AND ANTINEUTRINOS 611
where y = PM/~'M
and E
max
is the maximum neutrino energy (8 0).
For kaon neutrinos EK
max
= 0.95 Pm, whereas for pions, ETImax
A correction of (-7 2%) to the muon flux was made for "trapped"
muons from pion and kaon decays in the beam line (whose neutrinos miss
BEBC) on the basis of runs with the collimators of the momentum slit
closed.
TOTAL CROSS SECTIONS OF NEUTRINOS AND ANTINEUTRINOS 613
..
Q
N
~ 2
......
5 10 15 20
RADIUS em
/
.- 30m
""---
"'-
.....
2
1
-0
..~ \ --
10 40 20 o 20 40cm
w "Z
...J m
C> C
d EJectrons ( Messel 6 Z -t
<I :0
~ Crowford) Z
to- t- O
ToIol predidIed
; Z en
::::l }>
10
a: 0:: Z
...
0- W 0
a.. }>
..... ...... z
'"u" en ...... ..... -t
4 1
a: ::.:: ,, / ............ Z
to- \ u , ..... , m
... <I
\ '" 0:: E l eclrons C
...--t- - ...l- f ....-.... .
t- -t
:0
z
\\1r 'f, 0
I "-
en
I
I
1 0 1
50 100 200 300 400
ANGLE ""od ANGLE mrod
Fig . 6
0-
01
616 P. RENTON
- - --- ~ -. ---
E
"
.1<4.1. chambers
618 P. RENTON
The film was double scanned (efficiency ~ 99%) and the events
were measured using film plane digitizers and on-line geometrical
reconstruction. Figure 8 shows an example of an event. For each
event the following were measured:
I
I
\(/ .
,I
I
\. ~ "
"0 \ .
I
\ i
\ / .; . ".
1',! :\ ~ ~",
!, ~
\ . . ---""'-
" ~. . .. .
.'
\\ .,
. '. \.
NEUTRINO
E
u
.\ : \
\
lOO E
U 100 . \ .
rt: :.:
rt: .,
50 50 \ .,. :\
0
200 0 50 00 150 200
E GeV E GeV
Fig. 9
v,. + N ... r +
E
2
x= q 102M'"
~= -E v JL
E
where M is the nucleon mass and x and yare the (dimensionless)
scaling variables.
(2)
for \) (4)
for \i (5)
TOTAL CROSS SECTIONS OF NEUTRINOS AND ANTINEUTRINOS 623
q(x) 'teX)
SOc) :: (6)
q (X) ..,. ~ ( )()
G"v 1- B
R= - (7)
~ 2. + 8
cannot vary by more than 5% unless there are more than four quarks
or the W is unexpectedly light.
Table 1 -t
o
Neutrino and Antineutrino cross-sections i!r
C')
::n
11) o
Gargame11e BEBC (this experiment) ~
en
m
Energy Range Q
2-10 20-60 60-ll0 llO-150 150-120
(GeV) oz
en
V1f V1f VK VK o
aV/E 0.72 0.05 0.67 0.06 0.56 0.05 0.61 0.05 0.51 0.05
"Zm
C
136 123 161 97 -t
N - ::n
events
z
-
aV/E 0.29 0.92 0.26 0.03 0.25 0.03 0.32 0.04
oen
z
N - 90 87 73 o
events
z
- -t
Z
R = aV /av 0.40 0.02 0.39 0.05 0.45 0.06 0.56 0.07 m
C
-t
::n
JF2 (x)dx 0.48 0.04 0.45 0.04 0.39 0.04 0.43 0.04 z
oen
JXF3 (x)dx 0.41 0.06 0.39 0.08 0.30 0.08 0.24 0.08
Cross-sections (in units of 10- 38 cm 2 nuc1eon- 1 GeV- 1 ) are corrected to those for an isoca1ar
target. Errors quoted for the BEBC experiment are statistical. Systematic errors on all
cross-sections are estimated to be ~ 7%. Systematic errors drop out in the ratio R, except
for the v K point, where an uncertainty of 10% in the K+K- ratio should be included. ~
~
626 P. RENTON
Table 1 and Fig. 10 show the results for the neutrino and anti-
neutrino cross-sections. The low-energy Gargame11e data l l ) in freon
are also included for comparison 0
A ssum1ng avn/ avp = avp /0~n = 2,
small corrections have been applied to the Gargame11e results (3%)
and to the BEBC results (1%) so that all the data refer to an
isosca1ar target. The BEBC data incorporate other small correction
factors:
ot:,. GARGAMELLE
'" BEBC (this experiment)
S BEBC from muon spectrum
800.5
f Ii801.0
I
C\J
f I NEUTRINO
E
u 0.4
co
f")
'0 >----------.f..--------<I ANTI N E UTR I NO
0.2
-LdG'v
E d~
=0.15! D.09 0.71 ! 0.10 .
t'
I
0.6 t- R =o-il/o- 1/ -
t
0.5 - -
0.4 -9 -
0.3- -
o GARGAMELLE
0.2 f- BEBe -
0.1 ...... -
I I I
50 100 150
0.6
0.4 t
0.2
o
E GeV
Fig. 11
(dC'~!5 of d~/ci ~)
O. 79 :t O. f) 7 .
.tE.
This value corresponds to fF 2dx = 0.51 0.05 in Eq. (1). It is
determined from events of y ~ 0.05, and thus of q2 ~ 2 GeV 2 . It
can therefore be compared with the low q2 data from Gargamelle 11 )
1 3)
yielding 0.48 0.04; and with the SLAC e-d data via the quark
ed
model; (9/5)fF dx = 0.52. In summary therefore, the results
are consistent with both charge symmetry and the absolute values
of the neutrino and electroproduction cross-sections measured at
low q2.
TOTAL CROSS SECTIONS OF NEUTRINOS AND ANTINEUTRINOS 629
I I I I I I I
0,151-
w 0,10- 00 GARGAMELLE
-
ANTI NEUTRINO
"- X 15' FNAL
A
(\J SESG (this experiment)
c:r
V
I I I I I I I I I I I I
2 4 6 810 20 40 60 100
E GeV
Fig. 12
630 P. RENTON
the ratio q2/E = 2Mxy depends only on the muon track measurement,
and the total energy estimate enters only in the binning in E.
The BEBC data has been corrected for the p > 5 GeV/c cut, and pre-
~ 15)
vious results from Gargamelle and from the 15' FNAL chamber are
included. The lines indicate the empirical relation <q2>/E oc
O
oc E- .14 obta1ne
. d f rom an 1nt1re
. " 1y 1n"d epen dent ana1 0 f f1tS to
YS1S
" d " " b d 1 d . 15,16)
sca I 1ng eV1at10ns 0 serve 1n e ectron an muon scatter1ng
These results demonstrate again that there are significant devia-
tions from exact scaling q2>/E = constant) at neutrino energies
below 200 GeV.
/' Z
::0
// o
en
+ + - f- ./
lJ lJ (V) lJ lJ (V) o ././
~
./
./
observed 1 0
o
. 20 40 60
- background - 1.2 - .3
PfL+ GeV/c
------------------- >-------------- ------------
Net signal o 1.2 0 0.3
Fig. 13
0-
w
632 P. RENTON
8. DHmON EVENTS
REFERENCES
7) J.D. Bjorken and E.A. Paschos, Phys. Rev. 185 (1969) 1975.
9) C.G. Callan and D.J. Gross, Phys. Rev. Letters 22 (1969) 156.
17) D.J. Gross and F. Wilczek, Phys. Rev. D8 (1973) 3633 and
D9 (1974) 980.
MEASUREMENT OF NEUTRAL CURRENT CROSS-SECTIONS
H.P. Paar
CERN
Geneva, Switzerland
1. INTRODUCTION
v
II
(vII ) + Fe ~ nothing + anything
Here, "nothing" may stand for V (v) but we do not know. The talk
is divided into the following parts: neutrino beam, detector, da-
ta reduction, total cross-sections, energy dependence and y-
dependence of the cross-section.
2. NEUTRINO BEAM
At the beginning of this year the CERN 400 GeV Super Proton
Synchrotron (SPS) and its associated narrow-band neutrino beam came
into operation. The properties of this beam are especially impor-
tant for the study of neutral current interactions because the to-
tal energy of the final state is unobservable, contrary to the si-
tuation in charged current interactions.
635
636 H. P. PAAR
(1)
with
= EB/~'
-2
1 - YB '
energy of the V in the parent's centre-of-mass (30 MeV
for pions, 236 MeV for kaons),
EB parent beam energy (200 GeV),
~ mass of parent (0.140 GeV/c 2 for pions, 0.494 GeV/c 2 for
kaons) .
The neutrino energy and the decay angle in the lab. e are related
by
(3)
8000 r - - - - - - - - - - ,
RADIAL DISTANCE
40-80 em
6000
4000
2000
o '-'--__--".-o!=--_----'''----'
8000r----------------,
RADIAL DISTANCE RADIAL DISTANCE
80-120 em 120 -160 em
6000
4000
250
- Ev (GE'V) - Ev (GE'V)
Ul
::I
'1-1<II"'
<II
Po
Po
<II
III
.r:
- E-<
'~rT"~"""fI!.
N
.
MEASUREMENT OF NEUTRAL CURRENT CROSS SECTIONS 639
3. THE DETECTOR
4. DATA REDUCTION
4.1 Hadron energy cut
3500
3000
2500
2500
Jc:
~
w
1500
500
o
Event-time (lLsec)
1600,r------=~---------------------------------__,
Z1200
~
u
E
~
Cl:
w (a)
a. 800
(/)
.....
Z
~
W
LL
o
Cl:
~ 400
:::E
::l
Z I MONITOR REGION .. I
COSMIC MUON
/ BACKGROUND
600,---------------------------------------------.
Z
o
'!
E
u
~
Cl:
w
a.
(/) 300
.~
w (b)
>
W
LL
o
Cl:
w
CD
:::E
::l
Z I MONITOR REGION .. I
240
LENGTH OF EVENT (em IRON)
Table 1
Neutrinos Antineutrinos
The main flux of the wide-band beam background does not point
to our detector; still, this sort of background constitutes a
major source of statistical and systematic error in our analysis.
A fraction of approximately 5% of the running time has been de-
voted to closed collimator studies. The events observed in these
runs are scaled by the ratio of the collimator open and collimator
closed fluxes and subtracted. The subtraction is of the order of
4% for neutrinos and 29% for antineutrinos (see Table 1).
5. TOTAL CROSS-SECTIONS
The final result for the ratio of the inclusive neutral cur-
rent (NC) to charged current (CC) cross-sections on iron nuclei is
These ratios are integrated over neutrino energy; the average inci-
dent neutrino energy is 110 GeV for neutrinos and 90 GeV for anti-
neutrinos. In order to make a comparison with other experiments
and theory, we need to undo the hadron energy cut. For this we
need the hadron energy dependence of R = NC/CC in a region where it
is not measured (E H < 12 GeV). I will discuss this later, but now
already state the result:
,
12
I
0.8r
1.1
I
I
I
1.0
,I THIS EXP
o GARGAMELLE
I
0.9 Q7.j. b. HPWF
. I
CITF
I
0.8 I
I
1
0.7 +
Ry 1
\
0.6
0.5
0.4
0.3
0.2
01
6.1 Method
300 m
('")
:::!
cc o
z
Cf)
100
1"//#"//
A ,... I 0
100 200
(a) - EM (G eV) - EH (GeV ) (b) - EH GeV - EH (G eV)
04 08
02
03 1, l' t. 1I I j t .1 .. 0 6.
, T
,
.... _--- -
02 04h ~ "'~ I.YI
II 1 I I
01 02 'W
o 00 2000 100 200 o 100 200 0 100 200 ::c
(a) - EM (GeV) - EM (GeV) (b) - E. (G eV ) - E . (GeV)
~
Fig . 7 Ratio of NC to CC cross-sections as a function of the hadron energy for neutrinos (a) "'tI
and antineutrinos (b) . The solid line is the prediction of the Weinberg-Sal am model
:0
with sin 2 Sw = 0.24 . the dashed curve for sin 2 Sw = 0 . 35 . The dotted line represents
pure V-A .
MEASUREMENT OF NEUTRAL CURRENT CROSS SECTIONS 651
6.2 E
-\)
dependence of Neice
We fit the ratio R = Neice separately for E > 100 and E <
v v
< 100. The results are given in Table 2. So we have no evidence
for an energy dependence of R. It seems that the neutral current
cross-section is a linear function of E (because the charged cur-
V
rent cross-section is), a property required by scaling. From now
on we assume for simplicity that R is strictly independent of ener-
gy, although a small energy-dependence of 10% (for v) to 20% (for
v) cannot be ruled out.
Table 2
EV dependence of Rv and Rv
RV R_
v
6.3 y-distribution
gL + gR(1-y)2 + gsP y2
NC (v) (4)
CC
fL + fR (l-y) 2
gL(1-y)2 + gR + gsP y2
(5)
fL (l-y) 2 + fR
We find
0.02 0.07 .
NC gL + gR(1-y)2
(v) (6)
CC
fL + f R(1-y)2
gL (1_y)2 + gR
(7)
f L (1-y)2 + fR
We fit Eqs. (6) and (7) to V and V separately for the results
of Section 6.4 and 6.5. To increase the accuracy of the fits, we
MEASUREMENT OF NEUTRAL CURRENT CROSS SECTIONS 653
can make a common fit where gL and gR are required to be the same
in Eqs. (6) and (7). The latter is suggested by the quark-parton
model. We find
gL = 0.30 0.01
"common fit"
gR = 0.050 0.005
From Eqs. (6) and (7) we find for Rv and Ro (integrate numera-
tor and denominator over y):
3gv
V
L + gR
Rv (V)
3fL + fR
v + 3gvR
gL
R-V
fL + 3f R
RV 0.29 0.01
Rv 0.34 0.03
at y = 0
v + gRv
gL
(~~) y=o fL + fR
v
v
V + gR
gL
(~~)y=o fL + fR
v
Notice that Eqs. (6) and (7) appear to violate charge symmetry in-
variance for CC interactions:
0.30 r-..---,--,.,------,--,-----,----,.--,-----,
Ia = q /q =0.10 I
0.25
0.20
0.15
30- CONTOUR
(STATISTICAL ONLY)
I
0.10
V-A
1 5
a =
'2 - sin 2 8
W
+ - sin 4 8
9 w
5
b sin 4 8
'9 w
Table 3
V V V + V
+
V + Fe + ~ + X
We have plenty of these events, but we have also the wide-band beam
background (see Section 4.6). To eliminate it we make a cut:
Ev~s~ble
" > 100 GeV. There are no wrong-sign muons after this cut
out of 17,000 charged current (~-) events. There are 1100 neutral
current events with Ev~s~
, 'ble > 100 GeV. No event means less than
658 H. P. PAAR
8. CONCLUSIONS
3) The neutral current coupling is of the form (V, A), mainly V-A
(left-handed) but with a significant (four standard deviations)
V+A (right-handed) part.
Acknowledgements
DISCUSSION
CARTER:
amplitude NC N
Q
~
1M.,
~ 'Z. '2. I"!.
c.~ 0 ..,
"= w
u'2.
I~
1-2./ M~
M
where -1 ,
~= the m1x1ng angle (not necessarily equal to tan (~/~),
as in the standard model)
t~t+l)= the SU(2) Casimir eigenvalue of the Higgs multi-
y = the hypercharge of the Higgs scalars (plet
a = a numerical factor depending upon the definition of
the charge operator. In the standard model, ~ =2.
In this general model, the ratio of NC to CC amplitudes
is proportional to:
1'1'1
661
662 DISCUSSION
CABIBBO:
I could add a remark on this subject. If we assume
that the gauge group is SU(2)XU(1) and the usual assign-
ments for the quarks, i.e.: left handed particles in
doublets and right handed particles in singlets, the
neutral current jZ is determined in terms of aw , and
one is led to predictions for R~ and R~ of the form
R'I = f (~w). ( M.., )"""
~ ~C()\ew
R~ = ~ (Q..,) . (t-\.., \+
M2,cos9w )
fv and f~ are (under the above hypothesis) known function
DISCUSSION 663
HINCHLIFFE:
SANDA:
LIPKIN:
PAAR:
CHEN:
PAAR:
CHEN:
PAAR:
Missing energy does not cause a shift in hadron
energy, because the calibration constant makes up for it.
Of course, the resolution becomes worse. The calibration
constant is measured using a test calorimeter in a test
beam of pions and electrons with well defined momentum.
Electrons and pions give different calibration constants,
but the difference is less than 10%.
FERRERO:
PAAR:
This loss is due to failures in pattern recognition
and trackfitting. We have measured by hand 1000 events,
taken from the 10% lost events. Their distributions look
the same as the ones for events that are reconstructed by
the computer. So the loss does not introduce a bias. I
might add that we have also checked a few hundred events
that were successfully reconstructed by the computer and
remeasured them by hand. No systematic difference was
found.
KNOBLOCH:
FISCHER:
PAAR:
No; it was not taken into account. If we knew
something about it, we could take it into account. It is
possible, that this could raise the bound.
TAO:
So far, nobody has measured the direction of the
hadronic shower. This is not important in charged current
physics, since the muon angle is (very well) measured, but
in neutral currents it is impossible to get the x distri-
bution. Measurements show that the resolution on the
angle of a hadron entering a calorimeter varies roughly
like l/logE or a+b/E. Our Monte Carlo calculations show
that if the angular resolution is 10 mr at EH=lOO GeV,
then it is possible to distinguish between (1-x)3, and
(1-x)2 or (1-x)4 distributions. But if it is 20 mr,
then (l-x) and (1-x)5 look very similar. To compute it,
you need to know the structure of the hadron shower;
whether the hadron energy is distributed among one, two,
or more particles; how the energy is distributed among
them. This knowledge is crucial for the measurement of x
distributions. Could you give me some information on this?
PAAR:
We have information only on the lateral spread of
hadron showers from our test calorimeter. A preprint
describing this has come out.
ZICHICHI:
K. Winter et ale have measured the direction of
hadron showers in a test calorimeter and it seems to
work.
DISCUSSION 667
RENTON:
PAAR:
RESELE:
PAAR:
BUDNY:
PAAR:
SANDA:
PAAR:
The analysis of the energy dependence of the cross
sections is not finished, so we cannot say anything
about this at the moment. We will have this by the fall.
CHARGED V+A CURRENTS IN LEFT-RIGHT SYMMETRIC GAUGE MODELS*)
R. Budny
(2)
for small momentum transfers. Due to the factor 1.25, this is not
exactly V-A, but the consensus is that the currents of the usual
quarks are V-A analogues of the leptonic currents:
671
672 R. BUDNY
(3)
(5)
where jA and J V are the axial and vector parts of the neutral cur-
rent and IBi> is the bismuth nucleus.
(7)
lD3b D.()~)
(8)
lD.ID!: 0.0\")
674 R. BUDNY
(~ ,0)-1)
(10)
(ll)
and
and
+
The parameter b contributes only to the mass of W and can be R
chosen to make WR heavier than WL in order to make V-A dominate
over V+A. How much heavier WR must be will be shown below. The
parameters AL and AR contribute to the masses of the two ZO's.
With the (unnatural) choice AL = AR' the ZO's generate an effec-
tive neutral current interaction for electrons and nucleons which
is the sum of a jvJv term and a jAJA term. Each conserves parity,
so the parity violating effect in bismuth vanishes to lowest order.
(13)
where (V-A) is the sum of the currents in Eqs. (1) and (3). This
P
is manifestly left-right symmetric. If the WL and WR mix, we must
where
G = ~2.. r
(loS -$ i1"l -r') "Z. -t }z (c.o~! ",,"'ina -r Jz.
>li ~ T'Y112. - q '2. 8 m~-9'2.
(16)
and
1+tay..T
=
1 - tan!
1- 2\A-to '1A/>.
and
$; 1 + Zrt VA + 'lAA
i) !~!~~_E~!~~_!~E__~~~~Z_~~~_~_~~~~Z~
1 6)
The usual formulae for the rates continue to remain valid
if one makes the identifications:
(19a)
(19b)
( 19c)
(20b)
Here the suffixes F and G.T. mean transitions induced by the vec-
tor and axial-vector currents, respectively. (Pure Fermi and pure
Gamow-Teller in the allowed approximation). Also, the expressions
for the polarization are normalized in terms of the standard re-
16) .
su 1 ts der~ved ~n the V-A l~m~t.
17)
The standard parameters , p, 6, ~. and n. which determine
the spectrum are related to the interaction parameters as follows:
(2la)
(2lb)
s_ (2lc)
(2ld)
11,=0
(2le)
680 R. BUDNY
Table 1
Lower Upper
Quantity Experiment
bound bound
V-A
P(e)F/P(e)F 0.939 1 0.97 0.19 (a)
V-A
p(e)G.T./P(e)G.T. input 1 1.001 0.008 (b)
1"= 3- 4 r (22a)
31i+4p
(22b)
~&:r. ~&.'T.
< --;:::::===- (23)
1_ J1- ~&2..'T.
(24)
682 R. BUDNY
0.985~ PoT~ I
2.0
1.5
, ~~==========::~~~~~~~=========1-I .O ~AA
0 .994 ~ GI'GCO 9 !: 1.006
0.5
With P.
m~n
= 0.746, Eq. (24) gives the boundaries ab and cd in
Fig. 1. Note p cannot be greater than 0.750. The measurement of
~ and Eq. (2lc) constrains nVA and nAA to lie within the tilted
ellipse which is too large to give anything new. There is also a
theoretical constraint:
(25)
Ion ~
o b
005
,
0 b'
0 .1 0 .3 0.4 ~
0 m2
c'
d'
- 0 .05 d:r-----------------~ c
(26)
2V\ 2 -1 (27)
VA
Tab Ie 2
V-A
P(e)F/P(e)F 0.981 1.0
V-A
P(e)GT/P(e)GT input 1.0
and
(29)
When the momentum transfer _q2 gets large compared with m~,
the relations for rand s in Eqs. (22a) no longer hold. Instead,
r + 0 and s + 1 giving (to lowest order) the exact V-A results for
cc cc - 2
dcr (V L ) and dcr (v R ), except for a q dependent normalization.
CHARGED V+A CURRENTS 687
dcrt~t") - dcr(tA""l!: )
=
da- l\'t') +da- (f"p. )
(3:!.)
-
and similarly for dA- with (1-y)2 moved. Dropping terms of order
~2 and using S2 = (q2 - mf)2J{q2 - m~)2 + 8(~)2 gives
(32)
(33)
688 R,BUDNY
REFERENCES
5) T.P. Cheng and L.F. Li, Phys. Rev. Letters 38. 381 (1977).
6) H.A.B. Beg and A. Zee, Phys. Rev. Letters 30, 675 (1973).
F.A. Wilczek, A. Zee, R.L. Kingsley and S.~ Treiman, Phys.
Rev. D12, 2768 (1975).
H. Fritz~, M. Gell-Mann and P. Minkowski, Phys. Letters 22!,
256 (1975).
A. De Rujula, H. Georgi and S.L. Glashow, Phys. Rev. ~, 3598
(1975).
10) B.W. Lee and S. Weinberg, Phys. Rev. Letters 38, 1237 (1977).
P. Langacker and G. Segre, ibid. ~, 259 (1977).
12) J.C. Pati and A. Salam, Phys. Rev. DlO, 275 (1974).
R.N. Mohapatra and J.C. Pati, Phys.~v. Dll, 566 (1975);
ibid. 2558 (1975).
G. Senjanovic and R.N. Mohapatra. Phys. Rev. D12, 1502 (1975).
R.N. Mohapatra and D.P. Sidhu, Phys. Rev. Lett;rs 38, 667 (1977).
J.C. Pati, S. Rajpoot and A. Salam, preprint ICTP/76/ll.
13) M.A.B. Beg, R.V. Budny, R. Mohapatra and A. Sirlin, Phys. Rev.
Letters 38, 1252 (1977), and to be published.
6~ R.BUDNY
18) D.H. Wilkinson and D.E. Alburger, Phys. Rev. Ell, 2517 (1976).
. *)
N. Cab1bbo
691
692 N. CABIBBO
(1. 2)
where a lepton doublet has charges (0, -1) [example: (v , e-)] and
e
a quark doublet charges (2/3, -1/3) (example: (p, n)), in three
colours.
Here P., N. are the fields with definite mass (p, c, t ... , n,
~ ~
(2.2)
~ A
J+ = ~ 1>....). ~~')I'l
'V
). ,
L\C.
\.K
2. "- (2.3)
(2.4)
(2.5)
(2.6)
(2.4) and (2.6) show that the weak neutral current is diagonal for
any allowed (i.e., unitary) form of the mixing matrix A - the
theory is "natural".
696 N. CABIBBO
(2.7)
(2.8)
The general mixing scheme In the six-quark model will give rise to
T (and CP) violation 7 ) .
n A b
P C) -SlC 3 -S)S3
i6 i6
c SlC 2 C)C2 C3 - S2 S 3e C)C 2 S 3 + S2 C3e
A (2.9)
i6 i6
t SlC2 C)S2 C3 + C2 S 3 e C1 S 2 S 3 - C2 C3e
where
(2.10)
(2.11)
(2.13)
VV+!++Y.~
"e+e +X (2.16)
'X .
(-
v.,.~~- . . X,
\5e~Q"" ... 'X.
X
where ~ indicates normal hadrons, C a charmed particle. Such an
event could have up to four muons or three muons with signatures
-++ or -+-. Assuming equal semileptonic branching for Band C,
B ~ 1/10 we would expect:
~\iX
~1:2.:~ (2.17)
10
One would thus have roughly one 4~ event for each 30 events, or less
if alternative mechanisms for 3~ events exist (e.g., associated
charm production).
2.3 CP breaking
(2.18)
The reason for this result is that the first order (one W ex-
change) can easily be shown to be CP and T conserving. This is true
for any amplitude (including the neutron-photon vertex) which con-
serves all flavour quantum numbers. In a diagram for such an ampli-
+
tude, where the W ~s emitted by a (qlq2) term in J , it must be re-
absorbed by the corresponding (q2q 1) term in J
- The diagram will
then be proportional to A A* ,which is real.
qlq2 qlq2
Maiani and Ellis et al. (Refs. 7 and 8) have given an evalua-
tion of the second-order weak contribution, which leads to
QUARK AND LEPTON MIXING 701
3. LEPTON NIXING
p~e-+"'{
~ -. 3e.
(3.1)
\l-... N -+ e"," N
1 3)
The absence of ~ + ey at the value predicted by Feinberg on the
basis of the one-neutrino model prompted a set of theoretical in-
vestigations by Feinberg and Weinberg 14 ) , and by Gatto and myself 1S )
in which the merits of the three processes were compared, and dif-
ferent theoretical models were considered. We were led to conclude
that the only natural way to forbid the three processes was the
existence of a muon-electron symmetry which could only be realized
16)
through the existence of two separate neutrinos ,one coupled to
the muon, the other to the electron. This view was, as we now know,
correct. The advent and general acceptance of gauge theories has
allowed a sharpening of the theoretical predictions. The process
~ + ey can arise in some models at second order in perturbation
theory. The resulting amplitude is always computable in finite
terms, since in a renormalizable theory the only divergences are
those which correspond to terms in the Lagrangian, i.e., to possible
counter-terms, and gauge invariance forbids a direct ~ + ey coupling.
It has, in fact, been proved that any ~ + ey or ~-e (mass mixing)
term can be eliminated from a Lagrangian with minimal (renormali-
1 7)
zable) interactions by a redefinition of ~ and e fields .
For the other two processes, the situation is more complex and
we have to distinguish between two classes of models. In the first
class are models where lepton mixing can give rise to a ~-e-z coupl-
ing term in the Lagrangian. In these models ~ + 3e and the anoma-
lous capture process ~ + N + e + N can appear through a first-order
weak interaction:
QUARK AND LEPTON MIXING 703
j.l e I e
I
I
I
z Iz
I
(3.2)
I
I
I
I
e --------~-------- e q q
p .. all
(on S,
(3.7)
Ref. 22)
The results (3.4), (3.5) and (3.7) can be used to give limits
on a possible direct W-e-Z coupling. If we assume
(3.8)
-= (3.9)
IV 10
(3.10)
(3.11)
(3.12)
(3.13)
Ve.] 1.
2.
[('\-~')e ... ~"t-.)
(3.15)
, (3.16)
On the other hand, one easily computes the total rate for l decay,
which is proportional to the sum of the two mixings
(3.18)
-1 -9
We may safely assume rl < 0.3 10 s, which would give a mean
free path of ~ 10 cm in the SPEAR or DORIS experiments. If we
QUARK AND LEPTON MIXING 707
-4
'3 )\. 10 . (3.19)
The two bounds (3.17) and (3.19) together imply that either
of"
)
(3.20)
-r
q.
(3.21)
S2 y2 y2 S2
- + -
~ ~ ~ 'V 0 J%
-
e e e
+ -
3% 'V 0
- + -
~ e e 'V 0 2%
e
- ~ +~ - 2% 'V 0
708 N. CABIBBQ
CJ l..
(3.22)
In this case, neutrino (or charged lepton) mixing would not have
any physical consequence. Huon electron transitions are possible
if neutrinos are non-degenerate and mixed, but the corresponding
amplitudes will be proportional to mass differences, so that ap-
preciable rates are possible only if V is much heavier than V ,V
T e ]J
In this situation the result will essentially be independent from
the mass of the light neutrinos, so that we can assume them to be
massless.
)
e-) L
(3.24)
1 -9
Assuming, as in Section 3.3, a lifetime r~ < 0.3 10 we have a
(3.25)
(3.26)
'Z.
r'\-E
\ 2
) -VP. - abv~
The two neutrino states are not orthogonal, and not equally nor-
malized, so that:
-\)+w-.e--+
-.:>.. N~ p-.,. ... (3.28)
(3.29)
(3.30)
ve , VT
~
0 I
I
e
(3.31)
W I
I
I
I
I
Y
26)
One fInds
1<1=
QUARK AND LEPTON MIXING 711
(3.32)
27) . .
We h ave made a study of the relatIve merIts of ~ ~ ey and the
other two processes as a test of muon number violation. Within the
model under discussion, Fig. 1 shows the values of the three ratios
as a function of Mv = m, assuming sin 2 8W = 0.33 and (a b)2
-3 T
= 2 x 10 (equal to the limit in Eq. (3.30. In this computation
we have treated exactly the dependence on Mv
T
10- 6= - - - - - - - - - - - - - , Our results show that in left-
handed models, the anomalous cap-
ture is (for moderate values of
MVT ) 20 times more sensitive than
~ ~ ey. The situation is differ-
27
ent ) in other models including
right-handed currents, where ~
Fig. 1
712 N. CABIBBO
(3.34)
(3.35)
As was the case in the four-quark model (Section 2.1), one easily
sees that (3.33) gives the most general mixing among two neutrino
species, i.e., that any mixing involving phase factors can be re-
duced to the real form (3.33). If we have three neutrino species,
QUARK AND LEPTON MIXING 713
(3.36)
(3.37)
(3.38)
where
'Z..
-
m",
(3.39)
714 N. CABIBBO
(3.40)
(3.41)
(3.42)
(3.43)
(T conserved) (3.44)
1 1 1
(3.45)
A= 1- 1 Q 0.. t:l -
-
21r .
e.-,cp -3 '\.
~ tt
1 0.. 0..
QUARK AND LEPTON MIXING 715
This matrix has the interesting property that the three neutrinos
are maximally mixed (this isnot possible for three neutrinos with
a real matrix 29 )).
1
=-
(3.46)
(CPT) (3.47)
(CP) (3.48)
(3.49)
\J~_ v-r; VS, 'V.... ~ ~"t
The most convenient test for T violation seems to be the search for
odd (sin m2 /p t) terms ~n the oscillation probability.
REFERENCES
18) S. Parker, H. Anderson and C. Rey, Phys. Rev. 133 (1964) B768.
20) D.A. Bryman, M. Blecher, K. Gotow and R.J. Powers, Phys. Rev.
Letters 28 (1972) 1469.
24) D.A. Bryman and C. Picciotto, Phys. Rev. D11 (1975) 1337.
SANDA:
CABIBBO:
FISHER:
CABIBBO:
0-0 0-0
As in the K K system, CP violation can appear in D D mixing,
possibly with a strength comparable to that in KOK o mixing. However,
in contrast to the KOK o system, mixing effects are expected to be
0-0
very swall in the D D system since the leading decay channels
o -0
(having As =Ac = +1) will be separate for D and D. It is only
rare (proportional to sin 26 ) channels which are common to both
o -0 0-0
D and D decays and can lead to D D mixing.
One still expects the two states Dl and D2 to have definite
LIPKIN:
CABIBBO:
Yes, this would give a test but the expected combined branching
ratio would be sin~~ x (CP violating factor). This would be very
+ -
difficult to identify against e e ~4-r background!
DISCUSSION 721
DE LA TORRE:
CABIBBO:
L 1.. 2-
Not many. In the 3x3 matrix we introduced, the sum ~+S, Co 3 ::.
'1. '1. 1. 'L
'-I .... $, - $, S~ must be equal to 1.00 .05 in order to be
compatible with the data which justifies the old Cabibbo theory.
This doesn't put very strong constraints on S) however. It will
also be possible to measure the matrix elements which give the ci\
and cn couplings, but it will be a long time before this is done.
Of course we also have the relation E~SISl S~) where ~ ~s the
CP violating parameter of the K K system. If we believe the model
we expect the new angles to be comparable to the old one, perhaps
smaller.
PAAR:
You can see from the 3x3 matrix that it will be very hard to
make band t quarks because the production of valence quarks is
proportional to two sin~ type factors. So it will be very hard
to see band t quark production by neutrinos if this model is
correct.
CABIBBO:
CELHASTER:
CAEIBBO:
WEILER:
REITHLER:
CABIBBO:
If you believe the atomic physics data, there are many ways to
construct theories without parity violation in atoms. But you have
to keep in mind that the neutral currents observed in neutrino
reactions (either Jj.N or ~e ) agree quite well with the predict-
ions of the Weinberg-Salam model (in particular, the leptonic
neutral current seems not to be pure vector). If you accept the
atomic physics result and you wish to remain in the SU(2) X U(l)
framework, you can add a right handed doublet, (~)~ where N is
a neutral heavy lepton. This would give a purely vector neutral
current for the electron,and no parity violation in atoms (or to
a small violation, compatible with present limits). An alternative
to this approach is to enlarge the gauge group. Salam has recently
proposed an SU(2) X SU(2) X U(l) model which predicts no parity
violation in atoms but does predict parity violation in neutrino
induced neutral currents.
I wish to say that the point of my lectures was to illustrate
a possible model, not to show the absolute truth. The atomic
physics experiment will tell us something about physics, but what-
ever the results are it won't be a crisis for gauge theories if the
simplest model has to be modified.
724 DISCUSSION
BUDNY:
REITHLER:
REITllLER:
From the model discussed, an upper bound for the decay jt;er
has been predicted. How large is the predicted bound for the
analogous decay of the muon neutrino ?
'=rl-~~ I ~ - {[~/~J
DISCUSSION 725
this leads to
i;
where m is he mass of ~. So any experimental number for A and B
is interesting. How is A and B related to more fundamental para-
meters of the weak interaction theory? A and B can be computed in
a rather general way within the context of gauge theories, Consider
a Lagrangian
~ =- 4/
t!'
.2. ~,\ ~ ~ (A:j y.. . . .!. ij ... ) Lj
l j
C" = '!!)
+ Higgs contributions &
where >'i' stands for Y and ~ J L.is a set of heavy leptons with
mass ".... and charge Q which co~ple to both Y and YJot. We assume
')
that '\Nl<.<~ .
~J
M.~.#
~
where M~ is
~
the mass of the \Iboson.
A and Bare:
r
if the second term is non-vanishing
- 3D( (Mil"r)2.(~)1
r~~V'( -;r ~ Q r~
'2.
e>'Y lz.J
r~ . . .,r -- if 0(
-
2.~
,~
r
/,~e"~
(~)S.
INTERACTIONS
G. Preparata
There remains little doubt today that the quark idea must be
taken as the starting point of any serious attempt to build a theory
of hadrons and their interactions. But which direction one should
take appears at the present level of knowledge highly uncertain;
even though the successes of the gauge theories of weak interactions
have given a strong impulse to the belief that the "best" candidate
for a theory of strong interactions is Quantum Chromo-Dynamics (QCD) ,
which has the virtue of extending the gauge principle to the realm
of hadrons. Direct quantitative tests will decide which of the
proposed approaches, if any, should survive; for the time being it
is necessary to stress the importance that all possible avenues be
tried and confronted with the evergrowing experimental information
on the most diverse hadrodynamical aspects.
727
728 G. PREPARATA
ii) Quarks
From building blocks of hadronic symmetries [SU(2), SU(3), now
SU(4), and maybe SU(6) tomorrow] quarks have become important tools
first to classify hadronic states in the naive quark model l ) and
after as the elementary scatterers of deep inelastic phenomena 2 ) .
Thus the dynamics of the strong interactions from its low-energy
aspects (spectrum) to the highly inelastic phenomena (Bjorken
scaling) finds a powerful unification through the notion of point-
like s~in ~ constituents whose internal symmetry properties are the
same as the originally proposed quarks.
iii) Colour 3 )
DYNAMICS
+
GEOMETRY
Q CD + vacuum and surface tension Simple space-time domains
DYNAMICS
GEOMETRY
Spectrum and perturbative
Bags full of quarks and glue bag interactions
2.1 Simplicity
2.2 Freedom
(3.1)
i) Confinement
He require that
(3.2)
(3.3)
ii) Continuity
(3.5)
734 G. PREPARATA
(3.6)
The principle of freedom has not been ex~loited fully 1.n the
wave equation (3.6). Denoting by W(O) the solution of the free
equations
(3.7)
II +_
'\'\0) \\
(3.8)
1.S minimum.
'{e now briefly discuss the physical meaning of the four con-
ditions which characterize the discrete "fundamental" meson states.
or as
Max T<-"~~) .
)I.E: 1Z'c.~)
QUAR KGEOM ETRODYNAM ICS 735
!,.,.
2
-'
In fact were (3.5) not satisfied the momentum operators applied to
tpe (r)'s would yield infinity at the boundaries. Note that (3.5)
is the strongest continuity condition which can be imposed on the
wave function compatible with the wave equations (3.6).
(3.9)
(3.11)
(3.12)
(3.13)
(3.14 )
with
(3.15 )
and
(3.16 )
y:
tum, n radial quantum number, and are given by
~nf -~
- ~lt\l") ) (3.18)
th
Bn being the n positive zero of j(x). From (3.15) it also fol-
lows that
(3.19)
(3.20)
(3.21)
V! = [(y'+Yh.')'ifllPf+M J! ct>"'lf~X)
L )
(3.22)
QUARK-GEOMETRODYNAMICS 739
(3.24)
-+ -+
Decomposing (x,t) M(x)I/!H(t), as indicated by (3.14), (3.24) be-
comes
(3.25)
and
[(tl +~_A_M")L_ lMto- i)"] 4>", lb) = 0 J (3.26)
4 2
(3.28)
(3.29)
740 G. PREPARATA
We get
(3.30)
where
CUI = - "" + E1
2-
(3.32)
~2. = ~ - E:z
2
(3.33)
(3.34)
and combining (3.32) and (3.33) we can write the very simple eigen-
value equation for the mass of the state (nQ,):
(3.35)
where
(3.36)
+ +
742 G.PREPARATA
*) In order to overcome the famous U(l) problem one must assume that
there is some additional piece to the ninth axial current 12 )
QUARK-GEOMETRODYNAMICS 743
(3.40)
(3.41)
where, as usual, E
p,A
= ~2 + m2 ,.
P,/\
The requirement that the masses of the nseudoscalar octet de-
pend linearly on the chiral symmetry-breaking parameters (mp,mA),
as we shall see later on, through the s~ectrum equations, deter-
m~nes the low-mass behaviour of w(t!) as
(3.42)
where wo = 2\k 10 \' This has the consequence that the pseudoscalar
octet obeys a quadratic Gell-Hann-Okubo relation; ~n fact to first
order in m2 and m~:
p /\
(3.43)
2 'a.
Mp+mJ\.
(3.44 )
'2. '!c'o ,
z 'Z
""p ... 2",,,-
(3.45 )
6 '''10\
where the last equation follows from (3.42) by setting, as required
by the n-wave function,
lp) 2 l.\)
00,,= !OO'l~3~'l ,
where w(p) and w(A) are defined in (3.40).
n n
744 G.PREPARATA
All other meson states have no special role to play with re-
gard to the patterns of hadronic symmetries and their spectrum must
therefore follow the simple structures of Eqs. (3.35) and (3.36).
lve predict that all meson families are, before unitarity cor-
rections, in the ideal mixing configurations with the I = 0 non-
strange meson degenerate with the I = 1 member of the octet. In
Table 1 the calculated meson masses are reported with the following
parametrizations in Eqs. (3.38) and (3.39):
~: =4 G~\l2. (3.46 )
Table 1
~ 0 1 2 3 4 5 6 7 Particle type
5 s p o F G
6- 5- Firesausage band
7- 6-- 5-
7- 6- 5-- 4-- M= 4GeV
6-- 5-- 4---
4~6~~~~5~~~-'4~~~~----~~-
5 4 5-
5--- 4- 3-
4-- 3-- 2--
2---
3 3-- 2-
3-- 2-- 1-
3- 2-- 1-
1-
2-
2
2- 1--
1--
1--
1-
(3.50)
It should be noted that the mass of the D* is 1.93 GeV against the
experimental value ~ 2.0 GeV; but again shifts of this order of
magnitude are no cause for worry. Of particular interest is the
closeness of the vector states belonging to the subfamilies [(,n)]
(O,n) and (2,n-l). Such states as an effect of unitarity will mix
quite thoroughly, thus giving rise to mixtures of S- and D-waves.
Table 2
~ 0 1 2 3 4 5 Particle type
A similar situation obtains for the states (S,3) and (D,2), which
we predict at 4.28 and 4.24 GeV, respectively. They can account
for the bumps at 4.10 and 4.4 observed at SPEAR. Finally the mass
differences between D and F states come out around 100 MeV.
(3.51)
'YTl..a = S.l GftV ?
Table 3
The calculated spectrum of the T family (mu 5.2 GeV)
~ 0
9.53
1
9.80
2
9.97
3
10.10
4
10.21
Particle
T (H + H)
ty~e
(4.1)
(4.2)
i) Con.Nnement
(4.3)
:>
ii) Continuity
(r)
we require that the scalar functions abc(P;x,y) be continuous func-
tions, i.e.
(4.5)
(4.6)
(4.7)
(4.8)
(4.9)
-~
QUARK-GEOMETRODYNAMICS 751
MGeV
lP: 0- ,- 2- :r 4- S-
..2Q..
---1Q..
2i.22..
....2Q... -2L.2!L 22-
..2Q.. 22-
2L ..li..2!L 22-
..2Q.. A.2!L
..2Q..
-2L .iL2Q.. ..2Q..
2..22..
3 22- 22-
.iL .....2Q.. 22-
~2Q..
..1Q...
..:L
2 -2L -li.1Q..
..1Q...
~
1
Fig. 4 [SU(6),LP] pattern of physical baryon states
(4.10)
(4.11)
Table 4
The mean masses of the [SU(6),L P] multiplets
~ 1 2 3 4 5
(56, 6+)
2.91 3.48 3.94 4.34 4.71
(70, 6+)
(56, 8+)
3.27 3.81 4.25 4.64
(70, 8+)
(5.1)
(5.2)
(5.6)
where
and
754 G. PREPARATA
(5.7)
with
(5.8)
ti)- H
~-
2
(5.9)
1t"
.i)nl = (5.10)
~,ne
The wave function (p;x), in a general Lorentz frame in which
the meson is moving, can be obtained by (5.6), (5.7), and (5.8) by
the following identifications:
(5.11)
(p)(,) (5.12)
"1i\
and ~ 1S given by the direction of the spatial part of the four-
vector obtained from x by applying the Lorentz transformation A-1(p)
which brings the meson to its rest frame; i.e.
(5.13)
QUARK-GEOMETRODYNAMICS 755
(5.16)
(5.17)
and
(5.18)
(5.19)
(5.20)
++
where p = (l/2i) ca/'dx ). The actual form of the functions A and B
U ]J
is important only for the low-lying mesons; for high masses ~n
(5.21)
P,
P3
Fig. 5 The three-meson coupling
QUARK-GEOMETRODYNAMICS 757
(6.1)
= Z
,
N
V"
Sd\
~Jt)""
,.~ ()., y.
2. n
) (". ,,)
'In \: ) .
(6.2)
(6.4)
In the following we shall only kee~ the lowest vector mesons (V)
and write approximately
(6.5)
where
(6.6)
3
~ 3
= IK '3
~~/2
3
(jf3- m 3)
IK 23
(6.7)
(6.8)
the normalization factor NA is then determined by the equation
N"l. 1 (6.9)
A=
WA+ '}.A
In this way we have achieved a well-defined ~rocedure to nor-
malize the meson wave functions. After this is accomplished every
hadronic property can be computed starting from the basic parameters
of our theory [R~, c; mp ' rnA' mc ' ... , ].12J. In the next section
the first preliminary results on meson properties will be reported.
(7.2)
(7.5)
and we obtain
(7.6)
(7.7)
~ 1 JVVVV():;- 12
Fig. 8 The diagrams contributing to n(Q2) [Eq. (7.9)J
where
(7.10)
(7.11)
(7.12)
2:~.~
I (7.13)
I
I
(7.14)
This also looks quite good; however, one remarks that the effect
of the direct y-coupling as well as the finite mass corrections
have been left out. All these problems will be analysed in due
course.
By use of (7.2) we have for any radial recurrence of the J/W par-
ticle; W(n):
(7.18)
(7.19)
QUARK-GEOMETRODYNAMICS 765
(7.20)
with
(7.22)
and for K+
(7.23)
766 G. PREPARATA
1.1S
Experimentally, we have
l.mc)~ = (8.1)
defines the boundary between the meson states which have a wave
function of order 1 ( ~ ) and those which have an exponentially
max
o 10 20
M2 GeV
(8.2)
(8.3)
Rl (M)
~
..
~~ ~I~
~ no
2R2M
i) Scaling
1(. _ 2~
) - ~ J
(8.4)
and the function f(x,PT) is peaked for small values of PT' the mo-
mentum component transverse to the FS direction.
(8.5)
iv) CZustering
(8.6)
p (8) cc (1+C05 2 8)
It It
It It
(1) (2 )
At high energy, the process (1) can be shown to die off like
1/s2, owing to form-factor damping 9 ) and can be neglected. Process
(2), however, can be shown 9 ) to yield a cross-section which 1n-
creases like R~. One can, in fact, compute
(8.7)
(8.8)
This relation has been checked by Ferbel and Stix 23 ) and found con-
sistent with data; its correctness constitutes a good support for
the idea that in the geometrodynamical framework "strong inter-
actions" can be .treated as perturbations.
774 G. PREPARATA
mass cluster
2:n ~ _M 2
the new direction, thus yielding particles at large PT. The tilting
probability for a FS of mass M has been computed for scalar quarks
(see Ref. 9) from the diagram in Fig. 16, a similar calculation can
be carried out for spinning quarks with the result
(8.9)
where
\+* (.'+Gos&y"
l1-rA&)!.
and A} is a "coupling constant" whose value can be, with some fur-
ther labour, calculated. It is interesting to note that (8.9) is
similar, but not identical, to the t10ller scattering cross-section
for two quarks which has been invoked by people working with QeD
and the parton mode1 25 ). It has the same mass dependence (M- 4 ),
but a slightly different angular dependence, thus leading to results
which are similar to the parton model calculations. If we want to
compute the one-particle inclusive cross-section we must juxtapose
three separate pieces, as indicated in Fig. 17.
+~+
p
( tilting )
amplitude
leading cluster
leading cluster
2 dcr ~ A.
u: d3p ",4
T).
2E ~~ \
'ftO~
= A~ ~
n CJq
r)
dX,dy. a t ~ M."~,,
<Jc.,) t H ..c
-.
lx,) +
T
.te.wi..,., Q
rT
= Q (1p.
'(S)
e) .
[
n
where
where L 1 ,2 are the leading clusters and R(PT) denotes a low mass
(~ 1 GeV) resonance which gets emitted at large PT through one of
the two mechanisms analysed in the previous section. It is quite
obvious that an analytical description of such complicated con-
figurations is out of the question; thus we decided to follow a
strategy based on generating high PT events on the large CERN
QUARK-GEOMETRODYNAMICS 781
9. CONCLUSIONS
22) For a recent review, see the excellent lectures of B. Wiik and
G. Wolf, Proc. Ecole d'ete de physique theorique 1976, Les
Houches, France, Weak and electromagnetic interactions at
high energy (North-Holland, Amsterdam, 1977), p. 403.
24) As is well known, this fact has been taken as evidence for the
existence of gluons.
25) See, for example, S.J. Brodsky, Proc. 8th Internat. Symposium
on Multiparticle Dynamics, Kaysersberg, France, 1977 (C.R.N.,
Strasbourg, 1977), p. B-18l.
27) S.M. Berman, J.D. Bjorken and J. Kogut, Phys. Rev. D~, 3388
(1971) .
784 G.PREPARATA
CHAIRMAN: G. Preparata
DISCUSSION No.1
COLEMAN:
PREPARATA:
COLEMAN:
PREPARATA:
Yes.
785
786 DISCUSSION
COLEMAN:
So you are saying that all your solutions have the
same time dependence and the problem is effectively lin-
ear, and so there is no problem with superposition.
PREPARATA:
No problem.
HINCHLIFFE:
PREPARATA:
I view it as a coincidence.
HINCHLIFFE:
What do you do with all the new states described by
Perez-y-Jorba?
PREPARATA:
In fact I have sufficiently many states in this
region. Concerning the 1500 PP state I believe the 3
years old explanation of the Russian group of Shapiro
et ale giving a very narrow width and correct mass
might be the right explanation.
CELMASTER:
Are we to think of your wave functions as Bethe-
Salpeter wave functions? What is their meaning?
PREPARATA:
I will explain more about this in my next lectures.
For now, the wave functions contain information about the
momentum and space-time distribution of quarks inside the
hadrons but the details depend on the probe used.
DISCUSSION 787
BIGI:
PREPARATA:
CHEN:
PREPARATA:
REITHLER:
PREPARATA:
The masses come from the fit of the mass formula,
M = El + E2 - W with Ei =i cr,
particle spectrum of 'P, 6:>,
k 2 + mi to the physical
etc. The resulting values
of the mass parameters have very small uncertainties.
MOTTOLA:
PREPARATA:
?~:
is a non-local function ~i(P,x,y) unlike the usual local
vertices in Lagrangian field theory. This function comes
from our solution for the quark orbits in the bag. For
example,
~'~
corresponds to ).16JdX dy dz q>1(X,Z,P1) 'P2 (x,y,p,,).
CP~(y,z,PS) or ~Sdk ce(p. ,k) CP2(p:l.,k+P.:a /2 ) 'f3(P3,k-p:a./2)
~ {PI -p. -p) in momentJm'space. What we have is just
the ove~la~ integral for non-local vertices analogous to
local ~~field theory. Diagrams such as
correspond to
DISCUSSION 789
'3
in cp-theory and:
is analogous to:
-0-
with the non-local vertex functions substituted. This
last diagram is a self-energy correction or unitarity
diagram which must be perturbatively taken into account
to give mass shifts (real part) to the zeroth order mass
values and decay rates (imaginary part). If the frame-
work of the theory is reasonable, then these corrections
will be small.
CELMASTER:
PREPARATA:
BIZZARRI:
PREPARATA:
CELMASTER:
PREPARATA:
CELMASTER:
PREPARATA:
DEO:
PREPARATA:
DEO:
PREPARATA:
The interaction is simply proportional to the geo-
metric overlap of the wave functions of the initial and
final states.
DEO:
PREPARATA:
I have not seen their recent work but I believe that
their approach is very similar to mine, as they began the
work after my visit to Munich where I gave a talk on this
model.
BIGI:
PREPARATA:
BALDINI:
PREPARATA:
MIETTINEN:
What about the 1497 MeV narrow state and others ob-
served at Adone? You would not associate all of them
with <f' is that right?
PREPARATA:
LIPKIN:
PREPARATA:
BERLAD:
PREPARATA:
~ -'> s..e..n So
("Pomeron" exchange) is equivalent, as far as power
counting is concerned, to "gluon-exchange" in a spinor
theory. So the theory looks renormalizable, but I can-
not yet make this a quantitative statement.
BERLAD:
PREPARATA:
w'nen!. 1'2.
C1-x.L ")
CELMASTER:
PREPARATA:
1 _
DISCUSSION 795
CELMASTER:
PREPARATA:
JENNI:
PREPARATA:
DEO:
PREPARATA:
DEO:
PREPARATA:
DEO:
PREPARATA:
HINCHLIFFE:
PREPARATA:
BIGI:
PREPARATA:
BERLAD:
PREPARATA:
WIGHTMAN:
PREPARATA:
ATWOOD:
PREPARATA:
ATWOOD:
PREPARATA:
KRIPFGANZ:
PREPARATA:
MOTTOLA:
PREPARATA:
ATWOOD:
PREPARATA:
WEILER:
Your baryon wave functions contain a quark and a
diquark. So if you impose Fermi statistics on your sol-
utions the u and d quark components of the wave function
will have different spacial properties, leading to
F:n / t:..ep (X.1
\ r 6'I
'\) ...J. fT
as observed in experiment. Have you
tried to explicitly include Fermi statistics?
PREPARATA:
WETZEL:
PREPARATA:
WETZEL:
Have you started to think about neutrino scattering,
and coupling ~p. to hadronic states?
PREPARATA:
This just involves another axial current and won't
change the basic structure. It will just add another
kinematical form factor.
MARTIN:
Where do you get an entire function?
PREPARATA:
You get an entire function in the form factor be-
cause you cannot cut the quark lines. However for real
9~ the form factor could decrease very fast.
MARTIN:
However, unitarity corrections will produce a cut.
PREPARATA:
Yes, but this is a higher order effect. The point
like coupling is a bad thing. In deuteron it is well
understood because you have the two-nucleon threshold.
You can think of the situation here as if the q mass
goes to infinity. So you have a non-trivial 'i 2 behaviour
but no singularity. This is, I think, a very crucial
problem which everyone should be confronted with.
RAJPOOT:
Can one do weak interactions with your model? If
the answer is yes then does the idea of a firesausage
somehow turn into something equivalent to the inter-
mediate bosons?
802 DISCUSSION
PREPARATA:
Yes, it is possible to do weak interactions provided
one invents weak bags on similar lines to the theory
stated in my lectures. To implement this idea one would
have to alter the scale so that Ro goes down by a factor
of about a thousand.
WETZEL:
I come back to the SU(6) baryon multiplets: If I
understood you correctly the reason that the 20 dimen-
sional multiplets are absent lies in the condition that
the two quark coordinates have to be parallel. Now this
is the condition which has to be imposed for the solution
of the free approximation of the wave function. What
happens to the actual wave function?
PREPARATA:
The same condition also holds for the actual wave
function because it comes from the spatial part of the
differential equation.
WETZEL:
My second ~uestion is about the calculation of the
decay constant f~ , which is in good agreement with ex-
periment. The definition of f~ involves the pion mass
which presumably you cannot reproduce easily. Does
this affect f. ?
~
PREPARATA:
No. The value of m
TC.
drops out from the kinematics.
WETZEL:
Could you write down the qq wave function which
represents the state of the firesausage produced in
e+e- annihilation?
DISCUSSION 803
PREPARATA:
T.he wave function involves what I call a fat delta
function &~ j something which becomes a delta function
as R goes to infinity and it has a width of l/R. Then
~ = ZLo l2e+ l)
..eeO
~ (cps~) " ~ R:'....
z(pZ_m~") bo 2 C. ?,,~m~ )
1
e is the direction of the quark momenta with respect to
the firesausage axis.
WETZEL:
Should the angular momentum sum not be limited to
1= 1?
PREPARATA:
Yes, of course, but what this sum tells us is that
we must proj ect out the 1= 1. state. You produce your
firesausage as soon as your vector meson has emitted a
pion. The first step of the chain is determined by the
l=i projection probability but after that there is no
restriction on the angular momentum of the firesausage.
THE USES OF INSTANTONS
Sidney Coleman
I. INTRODUCTION
In the last two years there have been astonishing developments
in quantum field theory. He have obtained control over problems
previously believed to be of insuperable difficulty and we have ob-
tained deep and surprising (at least to me) insights into the
structure of the leading candidate for the field theory of the
strong interactions, quantum chromodynamics. These goodies have
come from a family of computational methods that are the subject
of these lectures.
These methods are all based on semiclassical approximations,
and, before I can go further, I must tell you what this means in
the context of quantum field theory.
To be definite, let us consider the theory of a single scalar
field in four-dimensional Minkowski space, with dynamics defined by
the Lagrangian density
' = g (1. 2)
805
806 S. COLEMAN
In terms of <p',
tJ.. = g\ (!- a]..l <P' a]..l <P' - ~ m2 <P' 2 - <P' If) (1. 3)
Thus, g does not appear in the field equations; if one can solve
the theory for any positive g, one can solve it for any other posi-
tive g; g is irrelevant. Another way of seeing the same thing is
to observe that, in classical physics, g is a dimensionful param-
eter and can always be scaled to one.
Of course, g is relevant in quantum physics. The reason is
that quantum physics contains a new constant, 11., and the important
object (for example, in Feynman's path-integral formula) is
L = ~ ~2 - V (x; g) , (1.5)
where
1
V(x;g) = 2"" F(g x) , (1.6)
g
Figure 1
decay. Only this last section reports on my own research; all the
rest is the work of other hands. 1
(Z.Z)
E In> , (Z.3)
n
then
<x Ie -HTft11 x. > = \' -E T/h <x In><n I x. > .
Len (Z.4)
f ~ n f ~
Thus, the leading term in this expression for large T tells us the
energy and wave-function of the lowest-lying energy eigenstate.
On the right-hand side, N is a normalization factor, S is the
Euclidean action 4
s = IT/Z dt ll(dX) 2 + vl , (Z.5)
-T/Z LZ dt J
and [dx] denotes integration over all functions x(t), obeying the
boundary conditions, x(-T/Z) Xi and x(T/Z) = xf . To be more
specific, if x is any function obeying the boundary condition, then
a general function obeying the boundary conditions can be written
as
x(t) = ~(t) + I cn x (t)
n
, (Z.6)
n
I T/Z
dtx(t)x(t)
-T/Z n m
<5
nm
(Z 7a)
x (T/Z) O. (Z.7b)
n
810 S.COLEMAN
oS (2.9)
ox
-= -
An x
n
(2.10)
v -v
------~~~-------+ X
( a) (b)
Figure 2
x= 0 (2.13)
(2.16)
(2.17)
-v
v -0 o
x
-0 o
(a) (b)
Figure 3
(2.l9a)
and
<a \ e -HT\ -a> = <-a \ e -HT\ a>, (2.l9b)
Figure 4
X
Equivalently,
t tI + f dx'(2V)-~ (2.21)
o
where tI is an integration constant, the time at which x vanishes.
This solution is sketched in Fig. 4; it is called "an instan-
ton with center at tI". The name "instanton" was invented by
't Hooft. The idea is that these objects are very similar in their
mathematical structure to what are called solitons or lumps,6
particle-like solutions of classical field theories: thus the "-on".
However, unlike lumps, they are structures in time (albeit Euclidean
time): thus the "instant-". For the same reason, Po1yakov suggested
the name "pseudopartic1e", also used in the literature.
Of course, we can also construct solutions that go from a to
-a, simply by replacing t by -t in Eq. (2.21); these are called
"anti-instantons".
Two properties of these solutions will be important to us:
814 S. COLEMAN
------~~-+--~L-~~------~t
T/2
Figure 5
That is to say, for small 11, the important terms in the sum are
those for which niT, the density of instantons and anti-instantons,
is exponentially small, and thus the average separation is enormous.
Note that this average separation is independent of T; our approxi-
mation is indeed a small-n approximation; the conditions for its
validity are independent of T, as long as T is sufficiently large.
This approximation of summing over widely-separated in-
stantons is called the dilute-gas approximation, because of its
similarity to the approximation of that name in statistical
mechanics.
(3) Finally, I want to deliver the promised fuller explanation
of the idea of an approximate stationary point of S. Let us begin
by studying an integral over a single variable,
I f dt e-S(t)f!r
T
(2.34 )
o
where S is a function of t monotonically decreasing to some asymp-
totic value, S(oo). Thus the integrand has no stationary points in
the region of integration. Nevertheless, it is easy to find the
approximate form of the integral for small~ and large T:
818 s. COLEMAN
(2.35)
(2.36)
dx = (dx/dt)dt i (2.37)
Hence,
(2.39)
<aie-HTi-a>
one inst.
(2.40)
where det' indicates that the zero eigenvalue is to be omitted when
computing the determinant. Comparing this to the one-instanton
term in Eq. (2.29), we find
det(-a~ + w2 ) ~
K (2.41)
det'(-a t2 +V"(X)
~
-1 o 2
(b)
Figure 6
THE USES OF INSTANTONS 821
As we see from Fig. 6b, the instantons are much the same as in
the preceding problem. The only novelty is that the instantons can
begin at any initial position, x = j, and go to the next one,
x j + 1. Likewise, the anti-instantons can go from x = j to
x j - 1. Otherwise, everything is as be fore.
Thus, when doing the dilute-gas sum, we can sprinkle instantons
and anti-instantons freely about the real axis; there is no con-
straint that instantons and anti-instantons must alternate. Of
course, as we go along the line, each instanton or anti-instanton
must begin where its predecessor ended. Furthermore, the total
number of instantons minus the total number of anti-instantons must
equal the change in x between the initial and final position eigen-
states.
Thus we obtain
1 2'IT
[ W )
'ITt!'
"2
e
-wT / 2
J e i (j - -'J+)8 21T
d8 -s "tf
exp[2KT cos 8 e ot]
o (2.44)
Thus we find a continuum of energy eigenstates labeled by the
angle 8. The energy eigenvalues are given by
(2.46)
v -v
Figure 7
~ t
Figure 8
THE USES OF INSTANTONS 823
I obtain the complete exponential series, rather than just the odd
or even terms, and I find that
(2.48)
(2.49)
-------'=--t--------'~ t
Figure 9
(2.50)
To show that this is the case requires a more careful argument than
Sagredo's. The essential point is Salviato's observation that the
energy of an unstable state is not an eigenvalue of H; in fact, it's
an object that can only be defined by a process of analytic continu-
ation. I will now perform such a continuation.
To keep things as simple as possible, let us consider not an
integral over all function space, but an integral over some path in
function space parameterized by a real variable, z,
(2.51)
Figure 10
v 5
___ _ _ _ _-+Z
x
(0) (b)
Figure 11
826 s. COLEMAN
Figure 12
Lie Algebras.
A representation of Lie a1gegra is a set of N
anti-Hermitian matrices, Ta , a = 1 ... N, obeying the equations
(3.1)
where the c's are the structure constants of some compact Lie group,
G. It is always possible to choose the T's such that Tr(TaT b ) is
proportional to cab, although the constant of proportionality may
depend on the representation. The Cartan inner product is defined
by
(3.2)
where the 0's are the Pauli spin matrices. In this case,
F
]lV
= a] l AV - aVA] l + [A ,A ] .
]l v
(3.6)
S = 412
g
fd4X(F
]lV
,F
]lV
) . (3.7)
where the A's are arbitrary functions. (Please do not confuse g(x)
with the coupling constant, g.) Under such a transformation,
(3.10)
and
(3.11)
A (3.12 )
]l
for some g(x).
Covariant Derivatives. The covariant derivative of the field
strength tensor is defined by
(3.13)
D F = 0 (3.14 )
]l ]lv
THE USES OF INSTANTONS 829
~ + g(x)~ (3.15 )
D~=a~+A~, (3.16)
]J ]J ]J
A -+ h A h- 1 + h d h- 1 (3.18)
].l].l ].l
Thus,
g -+ hg + O(1/r2) . (3.19)
(3.20a)
37T/2
o
Figure 13
THE USES OF INSTANTONS 833
og = i(oA)g , (3.22)
and the change in V vanishes upon integration. (We now know that
all of our standard mappings are in different homotopy classes and
that the winding number is uniquely defined.)
(5) If
g(8) (3.24a)
then
(3.24b)
(3.27)
Hence,
v=-i d x2 F
4rr ]lV ]lV
f
(3.28)
(3.30c)
V = 48~2
1 fd8 1 d8 2 d8 3 E ijk( gOi
'\ g-1 , gd g- 1 gdkg- 1 )
j (3.31)
THE USES OF INSTANTONS 835
(3.34)
(3.36)
this becomes
OV ~ Jd8 1d8 2d8 3E1]
k
Trdig-1djgdkoT , (3.37)
continuous deformations.
(5) Now to evaluate Eq. (3.32) for our standard mappings. The
task is easiest for g(l), for the integrand is here obviously a
constant, and we need evaluate it only at the north pole of the
unit hypersphere, x = 1, x. = O. At this point we might as well
It ~
d -1 .
(3.38)
g ig = -~oi '
and
(3.39 )
The argument is the same as for the baby problem, with semihyper-
spheres replacing semicircles.
(6) Let us define
G (3.41)
lJ
d G (3.42)
lJ lJ
From this expression we see that one way to make the surface term
vanish is to fix the tangential components of A on the surface.
lJ
Note that there is no need to fix the normal component of A . be-
lJ'
cause F is antisymmetric, this makes no contribution to the sur-
lJV
face integral.
We are not totally free to choose the tangential components of
AlJ arbitrarily. Firstly, they must be chosen consistent with our
gauge condition, A3 = O. Secondly, because we want to do semi-
classical computations, we must choose our boundary conditions to
be consistent with finiteness of the action, as the box goes to
infinity. Equivalently, the boundary conditions must be consistent
with the box being filled with a field configuration of a definite
winding number. Furthermore, for fixed boundary conditions, this
winding number is fixed, for only the tangential components of AlJ
are needed to compute the normal component of G. [See Eq. (3.4l).J
lJ
Thus at least one relic of our boundary conditions remains no
matter how large the box: we can not put an arbitrary finite-action
field configuration in the box, but only one of a definite winding
number. It turns out that the winding number is the only relic of
the boundary conditions that survives as the box goes to infinity.
The hand-waving argument for this is that the winding number is the
only gauge-invariant quantity associated with the large-distance
behavior of the fields. If you do not find this argument convinc-
ing, you will find a more careful one in Appendix D.
Thus, for large boxes, we can forget about the boundary condi-
tions in the functional integral and simply integrate over all con-
figurations where the winding number, v, has some definite value, n.
I will denote the result of such an integration by F(V,T,n). In
equations,
THE USES OF INSTANTONS 841
-S
F(V,T,n) = N J [dA]e 0 . (3.50)
Vn
where [dA] denotes [dA l ][dA 2 ][dA 4 ]. Also, I have set-h to one; we
can always keep track of the powers of -tr by keeping track of the
powers of g, as explained in Sec. 1.
F(V,T,n) is a transition matrix element from some initial state
to some final state (determined by our boundary conditions). ~fuat
This follows from Eq. (3.47), the expression for the winding number
as the integral of a local density; this tells us that the way to
put total winding number n in a large box is to put winding number
n l in one part of the box and winding number n 2 in the remainder of
the box, with n = nl + n2 (Of course, such counting misses field
configurations with significant action density on the boundary be-
tween the two sub-boxes, for there is no reason for the winding-
number integral for each sub-box to be an integer for such config-
urations. However, we expect this to be a negligible surface ef-
fect for sufficiently large boxes.)
Pretty as it is, Eq. (3.51) is not what we would expect from a
transition-matrix element that has a contribution from only a single
energy eigenstate. Such an object would be a simple exponential,
and would obey a multiplicative composition law for large times, not
the convolutive composition law of Eq. (3.51). However, it is easy
enough to turn convolutions into multiplications. The technique is
called Fourier transformation:
\' ine
F(V,T,e) - L e F(V,T,n)
n
f
N [dA]e- S e iVe (3.52)
From Eq. (3.51),
(3.53)
842 S. COLEMAN
of the 18> states without pausing to talk about the analogs of the
Ij> states, and that we did the Fourier transform that untangled
the energy spectrum before we saturated the functional integral
with instantons. The first difference is unimportant; if I had
wanted to, I could have added two extra paragraphs when I was talk-
ing about F(V,T,n) and discussed the analogs of the Ij> states.
(They're called n vacua.) As for the instantons, they are the sub-
ject of the next subsection.]
3.4 Instantons: Generalities
In the next subsection I shall explicitly construct instantons,
finite-action solutions of the Euclidean gauge-field equations with
V = 1. Most of the qualitative consequences of these solutions are
independent of their detailed structure and follow merely from the
fact of their existence. Therefore, in this subsection, I will
simply assume that instantons exist and draw some conclusions from
this assumption.
I will denote the action of an instanton by So' Because So
is finite, the instanton can not be invariant under spatial trans-
lations. Thus there exists at least a four-parameter family of in-
stanton solutions; I will call these parameters "the location of
the center of the instanton". The winding number is parity-odd.
Thus there must also exist at least a four-parameter family of solu-
tions with V = -1, the parity transforms of the instanton solutions,
which I will call anti-instantons. Just as in Sec. 2, we can build
approximate solutions consisting of n instantons and n anti-
instantons, with their centers at arbitrary widely-separated loca-
tions . These approximate solutions have V = n - n.
Again as in Sec. 2, we approximate Eq. (3.54) by summing over
all these configurations. Thus we obtain
(3.58)
I
<8 (F, F) 18> = -64n 2 iKe -So sin 8 . (3.59 )
Some comments:
(1) The expectation value is independent of V and T, as it
should be.
(2) The expectation value is an imaginary number, again as it
should be. The reason is that
(3.60)
(3) Both the vacuum energy density and the vacuum expectation
value depend non-trivially on 8. Thus the 8 vacua are indeed all
different from each other.
3.5 Instantons: Particulars
jd"x(F,F) [jd"X(F,F)jd"x(F,F)]~
~ Ijd"x(F,F)I , (3.61)
Iv I
811'2
S ~ -2 (3.62)
g
F = F (3.63)
A (3.64 )
11
where
(3.65)
A
jJ
-+ gA g -
. jJ
1 + 0 ( 1 /r 2 ) (3 . 6 7)
-S'
-2K'cos28e 0 (3.71)
where the primed quantities are the action and determinantal factor
for a binstanton. But S~ is twice So' so the new term is exponen-
tially small compared to the old one and should be neglected. 21
3.6 The Evaluation of the Determinant and an Infrared Embarrassment
We now know enough to go a long way towards explicitly evalu-
ating the right-hand side of Eq. (3.56).
THE USES OF INSTANTONS 849
(1) So is 8~2/g2.
(2) We have an eight-parameter family of solutions and thus
eight eigenmodes of eigenvalue zero in the small-vibration problem.
Thus K contains a factor of (1/~)8. or, equivalently l/g8. Every-
thing else in K is independent of~, and thus independent of g.
(3) He have already done the integral over instanton locations.
The integral over constant gauge transformations is an integral
over a compact group and thus gives only a constant numerical fac-
tor, the volume of SU(2). The integral over instanton sizes is
potentially troublesome, since p can be anywhere between zero and
infinity, so we will, for the moment, keep it as an explicit inte-
gral.
(4) Thus we obtain
J dp~
00
E(8) Iv = - cos 8 e
-8 ~21 g 2 g-8 f(pM) , (3.72)
o
where f is an unknown function and M is the arbitrary mass (more
properly, arbitrary inverse wavelength) that is needed to define
the renormalization prescription in a massless field theory. (I
have avoided mentioning renormalization until now, but renormaliza-
tion is essential in any computation that involves an infinite
number of eigenmodes, as does this one. In Sec. 5 I will give a
more detailed discussion of the ultraviolet divergences in deter-
minantal factors and their removal by the usual one-loop renormal-
ization counterterms.) The form of the integral is determined by
dimensional analysis; an energy density has dimensions of
II (length)" .
(5) However, M and g are not independent parameters. Renor-
malization-group analysis~ tells us that they must enter expressions
for observable quantities only in the combination
(3.73)
(4.1)
(4.2)
negative ~2, though: For positive ~2, the strength of the long-range
force is independent of~; for negative ~2, the strength of the
long-range force is exponentially small in~, the mark of an in-
stanton effect.
Just as in Sec. 3, we must begin the analysis by classifying
classical field configurations of finite action. Of course, before
doing this, we must add a constant to the Lagrangian density so the
minimum of the action is zero. Thus we write
(4.3)
This is the sum of three positive terms. In order that the third
term not make a divergent contribution to the action, it is neces-
sary that I~I approach a as r goes to infinity. However, there is
no restriction on the phase of~. In equations,
A (4.5)
~
Equivalently,
v =~
27T
fA~ dx~' (4.7)
fA~ dX~)
Wilson's loop integra1,27
W = exp (- ; , (4.10)
where the integration is over the rectangular path shown in Fig. 14.
According to Wilson, the vacuum energy shift is given by
THE USES OF INSTANTONS 855
TT'
1
Figure 14
In our case,
![dA] [d1jJ*] [dl/l]\.Je-Se iVS
<slwls>
J[dA] [dl/l*] [dl/l] e -S e iVS (4.12)
. inside
l>l = exp(Z1Tlq \! / e) . (4.13)
Thus, for the outside objects, we have the same sum as for the
denominator, except that the available volume of Euclidean two-
space is not LT but LT -L'T'. For the inside objects, we also have
the same sum, except that the available volume is L'T', and S is
replaced by S + Znq/ e .
Thus,
-S
.Q,n<SIHls> ZKe 0 [(LT -L'T')cos S
+ L'T'cos(S +Znq/e)
- LT cos S] , (4.14)
where the first term comes from the outside sum, the second from
the inside sum, and the third from the denominator. Hence,
where f, called the flavor index, labels the various triplets. The
usual exact and approximate symmetries of hadron physics [charge,
isospin, Gell-Hann' s sU(3) , etc.] act only on the flavor indices;
all physical hadrons are supposed to be singlets under the gauge
group. (This last statement is sometimes called quark confinement;
it is still far from proved, although there are some suggestive
arguments.) 1/J 1 and 1/J 2 form an isodoublet, the non-strange quarks;
1/J 3 is the strange quark; 1/J 4 is the charmed quark; there mayor may
not be additional flavors.
Chiral SU(2) SU(2) is the group generated by the strangeness-
conserving weak-interaction currents and their parity transforms.
Its diagonal subgroup is conventional isospin. This group is very
close to being an exact symmetry of the strong interactions; it is
a much better symmetry than SU(3) and roughly as good a symmetry as
isospin. However, were this symmetry to be exact, only the isospin
THE USES OF INSTANTONS 859
(5.3)
(5.5)
the relative minus sign in the propagators, and Goldstone poles ap-
pear where they should. This set-up is called a Goldstone dipole.
(The terminology is a bit misleading, because there are only single
poles in Green's functions, but I'll stick with it anyway.)
Thus according to Kogut and Susskind, the proper formulation
of our question is, is the D(l) symmetry of quantum chromodynamics
spontaneously broken via a Goldstone dipole? You might think that
this is a question that could be asked seriously only by a field
theorist driven mad by spending too many years in too few dimensions.
Nevertheless, as 't Hooft 9 brilliantly showed, the answer is yes.
The remainder of this section is an explanation of his computation.
5.2 Preliminaries: Euclidean Fermi Fields
Before we can treat quantum chromo dynamics by functional inte-
gration, we must know how to integrate over Euclidean Fermi fields.
This section is a description of the theory of such integration,
with all mathematical fine points ruthlessly suppressed.~ I will
develop the theory by defining Fermi integration as a "natural"
generalization of Bose integration. At the end, I will justify my
definitions by showing that they lead to formulas equivalent to
those obtained by conventional canonical quantization.
Let us begin by defining our integration variables. For Bose
theories, we integrate over c-number Euclidean fields. These are
objects that commute with each other at arbitrary separations; they
can be thought of as the classical (vanishing~) limit of quantum
Bose fields. This suggests that the proper variables for a Fermi
theory should be classical Fermi fields, objects which anticommute
with each other at arbitrary separations. Thus, for example, for
the theory of a single Dirac field, we would expect our integration
variables to be two Euclidean bispinors, ~ and ~, obeying
this were so, the last relation (multiplied by the inverse matrix)
would state that the sum of two positive semi-definite objects,
~~+ and ~+~, was zero. This would only be possible if ~ vanished,
not a happy situation for a prospective integration variable. Thus,
if we are to have any hope of founding a sensible integration the-
ory, we must treat ~ and ~ as totally independent variables.
This independence is the main novelty of Euclidean Fermi
fields; the rest of the construction is straightforward. We define
the Euclidean Y-matrices to be four Hermitian matrices obeying
{Y,Y}=20. (5.7)
)J v )JV
(5.8)
S = - Id"XijJ(iCl Y -
)J )J
im)~ . (5.9)
(5.15)
(5.17)
~ ==La ~
r r r
(5.18)
and define
[d~] [d~] IT da da (5.19)
r r r
As an apDlication let me evaluate
(5.20a)
where
S == - Jd 4 x~A~
-
, (5.20b)
A~ (5.21)
r
and we can choose ~ to be ,1,+. Thus
r 'I'r
S - rLAr a r a
r
(5.22)
and
J[d~] [d\jJ] e -S ITA
r r
det A (5.23)
Note that this is the inverse of the answer we would have obtained
had we done the identical integral with ~ and ~ complex Bose fields.
I will now show that Eq. (5.22) is the correct answer, that it
is identical to the normal field-theoretic expression for the
vacuum-to-vacuum transition amplitude in a theory of a quantized
Dirac field interacting with external c-number fields. In this
theory, this amplitude is the sum of all Feynman graphs with no ex-
ternal Fermi lines. This in turn is the exponential of the sum of
866 s. COLEMAN
Tr Ta Tb = _ C Gab . (5.24)
C = N/2 . (5.25)
f[d1/l][d\i!]e- S (5.27)
(5.28a)
(5.28b)
o ,
(5.29)
where jS is 1/IY Y 1/1.
]J ]J 5
These are, of course, just the Euclidean version of the Hard
identities we would have obtained in Minkowski space by studying
868 S. COLEMAN
the divergence of j 5, and, of course, they are wrong, for they take
"f.!
no account of the Adler-Bell-Jackiw anomaly. I don't have the time
here to recapitulate the theory of the anomaly, and I will simply
state the correct version of Eq. (5.29): The zero on the right-hand
side is replaced by
i C ( -) (1) (m) A
- 8n2 F(y),F(y) < (Xl)' . (Xm . (5.30)
Thus we obtain
. (1) (m) A
= - 4lCV< (x) ... (x (5.31)
I m
Now all our artillery is at the ready; we can begin our assault
on quantum chromodynamics.
5.4 QCD (Baby Version)
I will begin by analyzing a baby version of quantum chromody-
namics, in which the gauge group is SU(2), and in which there is
only a single isodoublet quark, of mass zero. In equations,
S = f ,,[l .-
dXL4g2(F,F)-l\)JD"f.!Y"f.!\)J
J (5.32 )
After we have worked out the baby theory, we will go on to the real
thing.
Most of the analysis of Sec. 3 is essentially unaltered by the
presence of a quark. In particular, all of our old instanton solu-
tions are still solutions of the Euclidean equations of motion
(with the quark fields set equal to zero). Thus we still have all
the 8-vacua, and formulas like
-S (3.56)
E(8)/V = - 2Kcos 8 e O '
and
THE USES OF INSTANTONS 869
[U~J
det i~J (5.33)
(A 0) (5.36 )
r
n -n =\!. (5.37)
+
Thus, not only is there a zero eigenvalue in the field of an in-
stanton, there is a zero eigenvalue in any gauge field of non-zero
winding number, whether or not it is a solution of the Euclidean
equations of motion.
870 s. COLEMAN
The proof rests on the chiral Ward identities for the quantum
theory of a massive quark interacting with an external gauge field.
(5.38)
(5.41)
Jd 4 y1jJ;'Ys1jJr 0 if A F 0 , (5.42)
r
while
Jd 4 y1jJ;'Ys1jJr = Xr if A 0 (5.43)
r
Thus,
- 2iv 2i(n+ - n ) (5.44)
cP
(1)
(xl)"'
A -
=
f [dl/J][dl/J]e
- -S cP (1) (Xl)'" (5.46)
[~+
oCt
2iV]CP (x ) .. A = 0 ,
I I
(5.47)
i.e., Eq. (5.31) without the mass term. The Green's functions of
our baby version of chromodynamics are given by
3
[ 3a + 2aeJ<81 (1) (xl) ... 18>
3]
= 0 . (5.49)
-iTIY
e 5 = -1 . (5.50)
where
(5.52)
(5.53)
K = 2g- 8 f (pM) ,
o
where M is the renormalization mass. As these old terms come up,
I'll call them to your attention, but I won't bother to write them
down; I'll keep explicit track only of new terms that modify the
integrand in Eq. (5.54).
There is one important novelty in the dilute-gas approximation.
For n widely separated instantons and anti-instantons, i~ has n
vanishing eigenvalues. Thus the integral over Fermi fields will
vanish unless the integrand contains
II a a (5 .55)
r r
A =0
r
Such a term can appear only if we are computing a Green's function
involving at least 2n Dirac fields. Hence, for any fixed Green's
function, the potentially infinite sum over instantons and anti-
instantons terminates.
I will first do the 0 computation:
In the denominator of Eq. (5.51), the only configuration that
does not have a surplus of vanishing eigenvalues is one of no in-
stantons and no anti-instantons, that is to say, the classical
vacuum, Ajl = o. Thus the denominator is simply the product of a
Bose determinant and a Fermi determinant. The same Bose determinant
874 S. COLEMAN
(5.56)
_8'IT2/g2
Finally, we have a factor of e from the instanton action,
i8 iv8
and a factor of e from the e
The a+ computation is almost identical to the a one; the only
difference is that the relevant configuration is one anti-instanton,
and thus, instead of a factor of e
is ,we have one of e
-is
Putting all this together, we find
00
(5.60)
and the fact that the only configurations that contribute to (5.62)
have \I = I, we find
~E: E: kn 1jJ. (l-y )~k~' (l-y )~n = 1jJ (l-y )~ ~ (l-y)~ -~ (l-y)~ ~ (l-y)~ .
1J IV 1 5 J 5 IV 1 5 1 2 5 2 1 5 2 2 5 1
(5.66)
These operators can have non-vanishing expectation values.
The doubling of C also changes Eq. (5.49) to
(5.67)
~ 1 , 2 -+ -i Y 5 ~ 1 , 2 ' (5.68)
(5.69)
we obtain
(5.70)
Figure 15
THE USES OF INSTANTONS 883
(6.2)
that goes from the false ground state at time minus infinity to the
false ground state at time plus infinity,
(6.6)
then,
(6.7)
Thus, at (j), 8 2S/82 has at least one negative eigenvalue, and (j) does
contribute to the decay probability. Of course, if there were more
than one negative eigenvalue, we would have to rethink the analysis
of Sec. 2.4. However, as I shall show eventually, this does not
happen; there is only one negative eigenvalue.
Now for the construction of the bounce: Eqs. (6.2)-(6.4) are
0(4) invariant. Thus it is not unreasonable to guess that the
bounce might also be 0(4) invariant, that is to say, that might
depend only on the distance from some point in Euclidean space.
Recently, Glaser, Martin, and I were able to show that this guess
is right, under mild conditions on U; there always exists an 0(4)-
invariant bounce and it always has strictly lower action than any
0(4)-noninvariant bounce. 42 The rigor of our proof is matched only
THE USES OF INSTANTONS 887
d 2 +1 Ei = u' () (6.11)
dr 2 r dr '
while Eqs. (6.3) and (6.:4) both become
lim
r-+oo
(r) = + (6.12)
Also,
(6.13)
-u
Figure 16
888 s. COLEMAN
(6.15)
U = U+ + E( - a)/2a , (6.19)
(6.20)
(6.21)
We can divide this integral into three regions: the outside of the
bubble, the skin of the bubble, and the inside of the bubble.
Within the accuracy of our approximation, in the outside region,
~ = ~+ and U = 0; thus we get no contribution from this part of
the integral. In the inside region, ~ = ~ and U = -E; thus from
this part of the integral we get
(6.23)
-a
S1 = I
a
Izu+ d~ (6.25)
(6.26)
(6.27)
Hence,
(6.28)
(6.30)
THE USES OF INSTANTONS 891
(6.31a)
and
(6.31b)
-+ -
(x o ,x) = (r = vlx
J'-+12-x 2 ) (6.33)
o
Figure 17
of 10- 10 _10- 30 sec. later, he is inside the bubble and dead. (In the
true vacuum, the constants of nature, the masses and couplings of
the elementary particles, are all different from what they were in
the false vacuum, and thus the observer is no longer capable of
functioning biologically, or even chemically.) Since even 10- 10 sec.
is considerably less than the response time of a single neuron,
there is literally nothing to worry about; if a bubble is coming
toward us, we'll never know what hit us.
(4) The rapidly expanding bubble wall obviously carries a lot
of energy. How much? A section of bubble wall at rest carries
energy Sl per unit area. Because any part of the bubble wall at any
time is obtained from any other part by a Lorentz transformation, a
section of wall expanding with velocity v carries energy Sll/l -v 2
per unit area. Thus, at a time when the radius of the bubble is
I~I, the energy of the wall is
(6.35)
894 s. COLEMAN
By Eq. (6. 34) ,
v (6.36 )
Thus,
(6.37)
fd 4 Xd (jid (ji
]l V
(6.38)
by Eq. (6.9).
Thus, as far as zero eigenvalues go, the only difference be-
tween the problem at hand and the particle problem of Sec. 2.4 is
!.:
that we have four factors of (So/2n) 2 rather than one. Hence,
I det
7z
[-d]ld]l+U"(<jJ~] (6.39)
V= fd 4 XU (6.40)
k
x det[-Cl Cl +U"()]2 (6.42)
].l ].l
Then
+ .. , , (6.44)
where the triple dots indicate terms that are negligible in the
order in which we are working. The second term vanishes because
THE USES OF INSTANTONS 897
f/V = 4:
S2
2 exp[-So - S(I) () +SCl) (~+)]
Appendix A
det(-a 2 +H)
t 2
-\jJ""'o-'(=T'::'/""'2):-- = 'IT 11 N (A.6)
(A.7)
Appendix B
(B .1)
+ Ae -I t 1 ,
The constant A is determined by the integral expression for the in-
stanton, Ea. (2.21),
t = rdX(2V)-~
= -~n[S:~A-I(a-~)]
+ O(a-x) (B.3)
o
Eauation (B .1) must have a second solution with A = 0, which I
denote by YI' It will be convenient to normalize YI such that its
l.Jronskian with Xl is given by
XldY-YdX (B .4)
t I I t I
Thus,
(B.5)
902 S. COLEMAN
Hence,
~o(T/2) = 1 . (B.7)
(B.10)
(B.ll)
(B.13)
THE USES OF INSTANTONS 903
As long as x is not near the bottoms of the wells, we can use stan-
dard yOCB solutions. Near the bottom of each well, though, there
are two turning points. These are not separated by many wavelengths,
so we can not use the standard connection formulas for a linear
turning point. Fortunately, near the bottom of a well, in a region
that includes both turning points, we may safely approximate V by a
harmonic-oscillator potential. Thus, for example, for x near a,
we may write
(B .15)
Ix
o
dx(2V)'2
1
fadx(2V)~1 -
o
JXdx(2V)~
a
= t So - Ha-x)2 (B.19)
For the E-dependent term in the integral, we can use Eq. (B.3).
Thus we obtain
1JJ (a-x) -~{ exp..tC 1 [t So - Ha-x) 2
+E R-n So-k2 A- 1 (a-x)]
x [l+O()] (B.22)
It will turn out that this is all that we need. Note that I have
normalized ~1 such that the Wronskian of the two solutions is
(B.25)
THE USES OF INSTANTONS 905
f dx' \jJ
00
\jJ = \jJ 1 1 (x' ) [\jJ 1 (x') <P 1 (x) - <P 1 (x') \jJ 1 (x) ]
- E (B.26)
x
I have chosen here the solution that vanishes as x goes to plus in-
finity. Thus, this is the appropriate solution for matching with
the decreasing WKB solution in the region (x-a) -ti. Thus, the
only matching left to do is in the region (a-x) ~.
f dx \jJ~
'-00
;.rr::ti (B.27)
to write
\jJ exp [ - (a-x) 2 /2tl:] [HO (E) ] - E ('Tf11:) \a-x) -1 exp [ (a-x) 2/ 2tr ] (B .28)
Appendix C
S =-tr/2 Lb n2
n
(C.2)
N'f[dXle-S/~ = 1 (C.4)
lim ; (2n)
L-><Xl n=l
-~f~ b n e -b~ o . (C.6)
-L
Q.E.D.
THE USES OF INSTANTONS 907
Appendix D
(D.3)
If we could choose
A = g d g-1 (D.4)
)J )J
A d -1
)J 3
)J g )J g ofo
= 0 , )J = 3 , (D.S)
is consistent with the gauge condition and will effect the transi-
tion.
We must compute the action associated with Eq. (D.S). If we
make a gauge transformation by g-l, Eq. (D.S) becomes
A
)J
o )J ofo 3
Appendix E
NO HRONG-CHIRALITY SOLUTIONS 47
F =
-
F (E .1)
~\! ~\!
o (E.2)
and
\jJ , (E.3)
is \jJ = O.
From Eq. (E .2) ,
o. (E.4)
Also,
Jd 4 XD \jJ+D \jJ
~ ~
= 0 (E.7)
Hence
D ,I, = 0 (E.8)
~ 't' ,
NOTES
1. These topics are all drawn from the classic part of the theory.
"Classic",in this context, means work done more than six months
ago. A good summary of the more recent research of one of the
most active groups in this field is C. Callan, R. Dashen, and
D. Gross, "Toward a Theory of the Strong Interactions" (to be
published in Phys. Rev.).
2. Polyakov's early work is summarized in A. M. Polyakov, Nucl.
Phys. B 121, 429 (1977).
3. See, for example, R. Feynman and A. Hibbs, Quantum Mechanics
and Path Integrals (McGraw-Hill, New York, 1965).
4. See the note on notation at the end of Sec. 1.
5. It was Polyakov (Ref. 2) who recognized the double well as the
prototypical instanton problem.
6. For a review of lumps, see my 1975 Erice lectures, "Classical
Lumps and Their Quantum Descendants", in New phenomena in Sub-
nuclear Physics, ed. by A. Zichichi (Plenum Press, New York,
1977) .
7. This is, of course, nothing but the standard prescription for
handling collective coordinates in soliton problems. See
J. L. Gervais and B. Sakita, Phys. Rev. D1l.., 2943 (1975).
8. The treatment here follows that of C. Callan and S. Coleman,
Phys. Rev. Dli, 1762 (1977). The idea of handling unstable
states this way goes back to Langer's analysis of the droplet
model in statistical mechanics [J. S. Langer, Ann. Phys.(N.Y.)
~, 108 (1967)]. The factor of ~, of which much is made below,
occurs in Langer's analysis and was explained to me by Michael
Peskin.
9. The order of my exposition will not be the historical order of
discovery. Here is the way it happened: The topological struc-
ture of finite-action Euclidean gauge-field configurations was
uncovered and the instanton solutions discovered by A. A.
Belavin, A. ti. Polyakov, A. S. Schwartz, and Yu. S. Tyupkin,
912 S. COLEMAN
28. S. Weinberg, Phys. Rev. Dg, 3583 (1975). This paper, titled
"The U(l) Problem", gives a characteristically lucid descrip-
tion of the situation just before 't Hooft's breakthrough.
(As a major unanswered question, Weinberg lists "How does the
underlying gluon-gauge invariance enforce the equal coupling
of the positive- and negative-metric Goldstone bosons to gauge-
invariant operators?")
29. S. L. Adler, Phys. Rev. 177, 2426 (1969). J. S. Bell and R.
Jackiw, Nuovo Cimento 60, 47 (1969). W. Bardeen, Phys. Rev.
184, 1848 (1969).
30. J. Kogut and L. Susskind, Phys. Rev. D1l., 3594 (1976).
31. For more details on Fermi integration, see F. A. Berezin, The
Method of Second Quantization (Academic Press, New York and
London, 1966).
32. The easiest way to fix this up is to stereographically project
Euclidean four-space onto a four-sphere; is then projected
into an operator with a pure discrete spectrum. This changes
the determinant, but only by a factor that is independent of
the gauge field. Since, as we shall see, our final results
will only depend on ratios of determinants, this change is
irrelevant.
33. To my knowledge, this sum rule was first derived by A. S.
Schwarz, Phys. Lett. 67B, 172 (1977). The derivation in the
literature closest to the one given here is that of L. Brown,
R. Carlitz, and C. Lee, Phys. Rev. D~, 417 (1977).
34. This section is mainly afterthoughts; I didn't know most of
these things at the time these lectures were given.
35. A related picture of how instantons break SU(2) ~ SU(2) is
advanced by D. Caldi, Phys. Rev. Lett. ~, 121 (1977).
THE USES OF INSTANTONS 915
43. See, for example, R. Jackiw, Phys. Rev. D2., 1686 (1974).
44. Formulas related to the one developed here can be found
throughout the literature. Two references out of many:
J. H. Van Vleck, Proc. Nat. Acad. Sci. 14, 178 (1928).
R. Dashen, B. Hasslacher, and A. Neveu, Phys. Rev. D 10, 4114
(1974). The derivation given here was developed in conversa-
tions with Ian Affleck.
45. This appendix reports on computations done with C. Callan. A
somewhat different attack on the problem (with the same con-
clusions) is E. Gildener and A. Patrascioiu, Phys. Rev. D16,
423 (1977).
46. I give no reference not because these results are novel but
because they are a standard part of the theory of Weiner in-
tegrals.
47. This appendix is a transcription of an argument of Brown et
al. (Ref. 33).
ACKNOWLEDGMENTS
DISCUSSION No.1
BUNK:
COLEMAN:
BUNK:
917
918 DISCUSSION
COLEMAN:
PAFFUTI:
COLEMAN:
)(
not only from finite parts of the plane where the inte-
grand is stationary, but if the derivative of the inte-
grand goes to zero as its variable goes off to infinity
then you'd better also worry about infinity.
TOWNSEND:
In the example of the double well potential you
computed the transition amplitude from +a to -a. This
required considering chains containing even numbers of
instantons. Had you commuted the amplitude for transit-
ions from -a to a you presumably would have had to
consider chains containing odd numbers of instantons.
Would you comment on how this changes the results?
COLEMAN:
You get, of course, exactly the same sort of thing.
Instead of the series for the sinh you get the cosh, but
still you get an appropriately weighted sum of two ex-
ponentials. And of course you get the right answer. In
this case you get
..... 1 (~ ):'i [E+T/i\
<
-q, t e
-+It:
\+(ll' = -
2.
-
11"'-
e - e'f..T/jt]
with a minus sign this time since the odd terms of the
series contribute instead of the even ones. The even
eigenstates contribute with the plus sign and the odd
with a minus sign. So, you know, whatever you want to
compute is fine: you always get the right answers. You
have to, the method is right.
MOTTOLA:
The n-instanton configuration functional integral
involves a determinant which is approximately the single
instanton determinant raised to the nth power. Is there
an easy way to see this?
COLEMAN:
Yes! I'll begin by making some definitions:
920 DISCUSSION
'Z ~
det--X2.t-(jt'+~+tt
"2. (
~-Tk
)
]
I obtain n..
= e){~ {_ ~ T'C' ..en l1 + ~ _~::~'Z f \+.- T" )) \ .
We all know how to evaluate this - it's just Feynman's
( "" ..... "-,1
perturbation theory with -"t+(A)) as the propagator
and f as an external source. So this is just the sum
of all the connected graphs
Tz, "l1 1
x----O---x + etc.. S .
n
de\:-'Yz [- '?>~ + ~ilt:-TJc.)+(b"Z.J = det:-~ [_~:;o(J)1.+ of (t) J
det:-y~ [- ~~ +fi] dQt-~ [-~t+~]
BOHR:
What is the difference between the Pontriagin index
and the winding number?
COLEMAN:
The Pontriagin index is just what I called the wind-
ing number. I called it the winding number in these
lectures just because I didn't want to introduce a lot
of terminology.
BOHR:
What can you do if your Lie group is not compact?
COLEMAN:
You mean if it's not simple. All gauge groups are
compact, so for our purposes we need only consider a
compact group. A compact group is locally a product of
factors - and it turns out to be only the local structure
that is relevant here - the product of a bunch of U(l)'s
and simple groups. For each U(l) there is no contribut-
ion to the Pontriagin index or winding number - no effect.
Every mapping of S3 to the circle can be continuously
deformed to the trivial mapping. For each simple group
in the product you have a winding number, and that wind-
ing number is in fact always computed the same way, by
the formula:
on:. __
1
1E.1t'2
5dx+ k- l FF)
,..,
.
WEEKS:
In classical Yang-Mills theory do you include in
your gauge group the transformations with winding number
different from zero or those that are disconnected from
the identity.
COLEMAN:
The full gauge group of the theory is the set of
functions from all of 4-space into the group. There is
no winding number attached to these mappings.
WEEKS:
In QED if you don't make some boundary condition at
infinity you find out that the Coulomb gauge is not a
gauge because the Laplacian is not invertible.
COLEMAN:
Physics determines the boundary conditions at infin-
ity. There's a physical reason why in QED when we work
in the Coulomb gauge we throwaway the solutions of the
Laplace equation that grow at infinity. Let me consider
the simplest case. Suppose the potential is a constant
times x, so that there is a constant electric field in
the x-direction. This is a perfectly legitimate solution
of Laplace's equation, and if you wish you can put it at
the beginning in QED. You can imagine someone has erect-
ed giant condenser plates at opposite ends of the galaxy
and charged them up so that there is a constant electric
field permeating all of space. Would you get a different
QED than you would get if you assumed that the electric
field goes to zero at infinity? The answer is no, becaus4
as soon as you couple in the electrons you make the vac-
uum polarizable, and the vacuum suffers dielectric break-
down. Electron-positron pairs will be pulled out of the
vacuum, and if the condenser plates are far apart, they
get a lot of energy as they travel t'o the plates, more
than enough to compensate for the 2mc 2 required to make
them. The electrons would fly to the condenser plates
until they were neutralized; and, after this enormous
DISCUSSION 923
JANCEWICZ:
I did not understand yourexample with the two gauge
fields with the same winding number. You said that these
two fields are connected by a gauge transformation, but
the gauge transformation you chose is the identity at
the origin.
COLEMAN:
COLEMAN:
Firstly, although it is not perturbative, this is
a systematic weak coupling expansion. Thus, all argu-
ments you make about the orders of things are the same
as you would make in perturbation theory. In particular,
all renormalizations occur just as in perturbation theory.
For example, though the determinant K is formally ultra-
violet divergent, that divergence is cancelled by the
standard counterterms, and the finite parts of those
counterterms are determined by the renormalization con-
ditions. There is reference to the renormalization
924 DISCUSSION
PAFFUTI:
How good an approximation is it to compute the eff-
ect of tunnelling with a gas of noninteracting instantons?
COLEMAN:
Very good, as long as the coupling constant is very
small, and there are no long range forces between the
instantons. Suppose the instantons have a definite size,
so we can ignore the integration over instanton scale.
Then we are doing a sum of the form
_ sol1\. \Y\
L.!.le KVT)
n. n!
The dominant contribution to the sum comes from the terms
for which
01'
KENWAY:
COLEMAN:
I have no profound insight into their computation.
I look at it and say, "Who would believe that?" But, on
the other hand, I say, "Is it evident nonsense?" I prefer
to say nothing at all about it here.
As for the fact that instantons will not give rise
to confinement in four dimensions, that's obvious. We
do the loop calculation again. Instantons near the
boundary of the loop can give rise to a large effect, but
they contribute a factor like ~X~(L+'). That doesn't give
us a linear potential. We need a contribution from the
instantons inside the loop. We might naively expect to
get one, because the gauge fields far from an instanton
falloff like l/r. But the part of the instanton that
goes as l/r is a pure gauge, and, in four dimensions, I
can make a gauge transformation such that the gauge field
vanishes outside a small tube perpendicular to the loop,
so it doesn't contribute at all. There is no hope of
getting a contribution from the instantons that goes as
e)(p- c..A~e.o. ') .
't HOOFT:
COLEMAN :
But that's completely different. Polyakov does not
use the dilute gas approximation. He studies the Georgi-
Glashow model, in which the instantons are the
926 DISCUSSION
WIGHTMAN:
Is it not true that the probability for a field to
approach infinity at r = 00 is finite when measured in
the probability measure defined by the action? (In fact,
I know that for a free scalar field the answer is yes;
see the article by Lanford and Collela in the 1973 Erice
Mathematical Physics School.) Thus, the probability
that the action is finite is zero. It is remarkable
that in spite of this, the points in function space where
the action is finite and stationary seem to dominate the
integral. Do you have any comment on this situation?
In particular, if one writes
r -~/-n.
J [dA1 e. 0\- " 'Rest"
n-\'ilted
S -ti~ite
do you have any argument why "Rest" =0 ?
COLEMAN:
I know of no such argument. But the problems arise
just as much in one-dimensional quantum mechanics as they
do in field theories. If they cause a disaster, they
should cause a disaster in the potential problem consid-
ered in the first lecture. But they don't. You get
exactly the same answer using the functional integration
methods as in good old wave mechanics using WKB.
Also, when you look at one-loop corrections in four
dimensions, of course, you get infinities, but they are
cancelled by the usual renormalization counterterms. The
renormalizations work out exactly right using the funct-
ional integration. So, at the level, not of a quantum
field theorist, but of a quantum field butcher, the level
at which I operate, so far, I would say, things look good.
't HOOFT:
COLEMAN:
Yes. One can observe these negative powers in my
own treatment of the potential problem. Remember that
I found a factor of r.
Soli\.1~ which came from the change
of coordinates that eliminated the translation mode. In
four dimensions, this would be raised to the fourth power,
and additional powers come from zero modes associated
with rotations in the gauge group, as you point out. I
have no comment, except that this is a clever idea about
the source of the infrared cutoff.
't HOOFT:
COLEMAN:
I don't know. I do expect there to be charged part-
icles in the theory because of the [1- Cd..l21Tq/e)] factor
in the force. This vanishes for q an integral multiple
of e, which I can only explain by the mechanism I ment-
ioned, the breaking of the string. So there are part-
icles of charge e which are confined in a linear pot-
ential which is very shallow, because of the ~~GS~~)
factor. Hence, I expect a large number of closely spaced
neutral bound states, as in the massive Schwinger model.
Whether some of these bound states can be identified with
the Higgs meson or the massive vector meson, I don't
know. I do know that this rich spectrum of bound states
is not what I would get by applying perturbation theory
to the Abelian Higgs model.
928 DISCUSSION
't HOOFT:
COLEMAN :
TOWNSEND:
COLEMAN:
then
~ = -1 (F;" y~ + B~ .
I now have a theory of a massive vector meson. But no
one presented with this final theory could possibly tell
it was a result of the Higgs phenomenon; it's just the
theory of a free massive vector meson. Unlike the
Goldstone phenomenon, the Higgs phenomenon leaves no
footprints. In fact, in simple models that you can anal-
yze, it's very easy to prove that in the gauge Ao=o
the vacuum expectation value of the scalar field is zero.
In a covariant gauge, of course, it can have any value
around the circle. So it's a bad way to ask the question
to say~"Do the instantons destroy the Higgs phenomenon?"
You can destroy the Higgs phenomenon by adopting a diff-
erent gauge.
JANCEWICZ:
COLEMAN :
PAFFUTI:
COLEMAN:
gauge, <
\C/.I.(~) <r(o) >
phase factor, so the theorem states that, in a covariant
must have a Goldstone pole; that
is, a pole at zero mass in its Fourier transform. Indeed,
if we do a sum over instantons to compute it, we find that
<
k"C.x.) crlD) does have such a pole. However, in the
same gauge, {<rtX.')()(D' and <k... Cx.) lC.olo)) have no
Goldstone pole. The first statement tells us that there
is a zero mass particle which ( j makes from the vacuum
and K" annihilates. The other statements seem to say
that this is not so. What's wrong?
The problem is that we are working in a covariant
gauge, where there are lots of negative norm states. To
see how precisely this structure can arise, consider a
theory in which I have two mass zero scalar fields
r----,
LIPKIN:
COLEMAN:
't HOOFT:
COLEMAN:
PRESKILL:
COLEMAN:
TOWNSEND:
COLEMAN:
CELMASTER:
COLEMAN:
It HOOFT:
COLEMAN :
CABIBBO:
COLEMAN:
WHITE:
COLEMAN:
BUDNY:
COLEMAN:
KENWAY:
COLEMAN :
{(it
If you take a typical microphysical energy difference,
for example, the difference between the energy of the
symmetric and asymmetric vacuums in the Weinberg-Salam
model for some randomly chosen Higgs meson mass, you find
that this number is of the order of one kilometer. So
we do not have to worry about gravitational effects dur-
ing the formation of the bubble, a quantum process, but
during the later stages of the evolution of the bubble
you might expect gravitation to become important. In
fact, gravity is not important in the expansion of the
bubble. There are two arguments to support this, and
Frank De Luccia and I have done an actual computation to
back this up. The first argument is as follows. Assume
for simplicity that the catastrophe has yet to occur, so
that outside the bubble, space is essentially flat. Con-
sider how the expansion of the bubble looks to an obser-
ver outside the bubble wall. I have already argued that
gravitation is unimportant during the early stages of the
bubble expansion, so during this time the bubble wall
describes a hyperboloid. The equations of motion are all
Lorentz invariant - the gravitational field equations do
not break Lorentz invariance. Since the equations are
0(3,1) invariant and the initial conditions are 0(3,1)
invariant, so is the solution, so the expanding bubble
wall is still a hyperboloid. The second argument is to
appeal to Birkhoff's theorem which says that the gravit-
ational field outside a spherically symmetric expanding
or contracting object depends only on the total energy
within it. Now the total energy of the bubble is the
sum of two terms, the positive energy of the bubble wall
and the negative of the interior. As I have shown,
DISCUSSION 939
MOTTOLA:
COLEMAN :
MOTTOLA:
COLEMAN:
DE LUCCIA:
COLEMAN:
TOWNSEND:
COLEMAN:
G. 't Hooft
1. INTRODUCTION
943
944 G. 't HOOFT
combination
The whole complex plane is now mapped onto the interior of the
unit circle, with the singularities at the edge. If we rewrite
the perturbation expansion in terms of u instead of g then we
have convergence everywhere inside the circle*), which implies con-
vergence for all g2 not too close to the negative real axis. A
definite improvement. Unfortunately, the structure for complex g2
we find in Section 4 is too complicated for this method to work:
the origin turns out to be an essential singularity. Still, I
shall show how this knowledge of the complex structure may be used
to give a very slight improvement in the perturbation expansion
(Section 9).
(2.1)
Here we show explicitly the minus sign for the first coefficient.
This minus sign is a unique property of non-Abelian gauge fields
and is responsible for "asymptotic freedom" (at increasing ].l we
get decreasing g2, see Refs. 14-16).
21-23)
The coefficients 81 and 82 are known . Since we shall
try to go beyond perturbation expansion we must be aware of two
facts: First, the perturbation expansion is expected to diverge
for all g2 and is, therefore, at this stage, meaningless as soon
as we substitute some finite value for g2. Second, the dimensional
procedure has only been defined in terms of the perturbation ex-
pansion (Feynman diagrams). Consequently, g~ may not have any
meaning at all as a finite number. The correct interpretation of
these series is that they are asymptotic series valid for infini-
tesimal g only, or, equivalently, valid only for asymptotically
large moment: p2 O(].l2) + 00. Thus, g~ may not be so good to use
as a variable for a study of analytic structures at finite complex
values.
(2.2a)
(2.2b)
948 G. 't HOOFT
The series in (2.2b) must stop after the second term. In pertur-
bation theory these requirements have a unique solution for g~.
For instance, we get that the rest term in (2.2a) is
(2.3)
(2. Sa)
and
(2.Sb)
We emphasize that the new parameters gR and ~ are better than the
previous ones, because for any theory for which the perturbation
expansion is indeed an asymptotic expansion, they are completely
finite and non-trivial. Of course, they still depend on the sub-
traction point~. At infinite ~ they coincide with other defini-
tions; at finite ~ they are finite because we can solve Eqs.
(2.2b) and (2.Sb):
(2.7a)
(2.7b)
(3.1)
. f y a renorma1
They sat1s
1zat1on . 24) :
group equat10n
(3.3)
or
(3.5)
chosen, that is, on the constant ~o, which we get in solving (3.3),
see Eq. (2.7a). Thus we get
(3.6)
with
(3.7)
and for dimensional reasons, G(k2,~o) can only depend on the ratio
k2/~~.
~R
'Flo) '1'(0)
(4.1)
(4.3)
-/. ..:-- o
-I
5. BOREL RESUMMATION
(5.1)
(5.2)
then
(5.3)
(5.4)
=_is'(A') (5.5)
~1
g'(A'): iCA,)L+V(A')
2.
Our integral becomes
(5.6)
(5.8)
CAN WE MAKE SENSE OUT OF "QUANTUM CHROMODYNAMICS"? 957
'JS'(A) = 0 (5.9)
7JA
to be recognized as the classical field equation of this system.
At a solution Aof (5.9) we have
%=S'(A)
[ C>S'] -1 [2 ';)2.SI ex _%)]-of/2- (5.10)
"i>A ~ C1A2..
Q.s z-+ x
;) S'
~=O (5.11)
'JA
because those are the contours that shrink to a point (in the case
of a local extremum) or have crossing points (saddle points).
Again, Eq. (5.11) ~s nothing but the classical Lagrange equation
for the fields A. Conclusion: to find singularities in F we have
to search for finite solutions of the classical field equations in
Euclidean space-time. Their rescaled action S' corresponds to
singularity points ~ in the z place for the function F. In general,
these singularities are branch points. In our actual four-dimensional
958 G. 't HOOFT
field theories that are supposed to describe strong (or weak and
electromagnetic) interactions; such solutions indeed occur, and are
called "instantons", because they are more or less instantaneous
and local in the Euclidean sense 25 ,26,6,lO). Their action (in the
case of QeD) is S' = 8n 2n, where n counts the "winding number"a-lo)
and so we may expect singularities in the complex z plane at z =
= 8n 2 n.
Let
(6.1)
Then, if
CAN WE MAKE SENSE OUT OF "QUANTUM CHROMODYNAMICS"? 959
then
(6.2)
And let
then
(6.3)
Here the symbols are there just to tell us to l~ave the 0 symbols
out of the integration. It is easy to show that if (5.3) is solved
iteratively, then the series converges for all z, as long as Fl
stays finite between and z.
Now note that we may choose the contours (O,z) so that they
avoid singularities. Only if z is a singularity of either Fl or
F2, or both, then F 3(z) in Eq. (6.2) will be singular. Also F2(Z)
in Eq. (6.3) is only singular if F (z) is singular. Note, however,
that if a singularity lies between 0 and z then the contour can be
chosen in two (or more) ways, and, in general, we expect the out-
come to depend on that. Thus if we start with pure pole singulari-
ties, they will propagate as branch points in the other Borel func-
tions.
(7.1)
we have
(7.2)
(7.3)
(7.4)
2. 't1 )( \)
Au = ~ A
0.' Q.
='OCf(~ (7.5)
,- f4- )( 2. + e,1
. 6,27)
where n are certa~n real coeff~c~ents and p ~s aga~n a free
a~JV
scale parameter. One finds for the action
(7.6)
(7.7)
(7.8)
(7.9)
8. OTHER SINGULARITIES IN F
(8.1)
(8.2)
where a is some fixed power. After having made the necessary sub-
tractions to make the integral converge, and in order to obtain
physically relevant quantities, such as a magnetic moment, the
leading coefficient a becomes 3 or larger. Let us replace log k 2
by a new variable x, then (8.2) becomes proportional to
(8.3)
(8.4)
(8.5)
(8.6)
?
-6 -3 4 6 8
instantons renormalons
The situation for QeD is more complex. Not only do we have the
renormalons at points on the negative real axis but also there are
such singularities on the positive real axis. They are due to the
infra-red divergence of the theory. The mechanism is otherwise the
966 G. 't HOOFT
?
1 2 3 4
renormalons
IR divergencies
?
)) instantons
+
renormalons -4
lr,
(9.1)
A x/2
WCs) = LCO F(x.) dl.. S'-'" (9.4)
We may hope that this has a finite region of convergence, from which
we can analytically continue. Note the analogy between (9.1) and
(9.4) on the one hand, and (5.2) and (5.3) on the other. The inte-
gral relation between Wand G, analogous to (5.1), is
(9.5)
(9.6)
(9.7)
Now we can easily prove that, if our theory makes any sense at all,
there may be no singularities in W(s) on the positive real axis, ane
the integral (9.7) must converge rapidly. Thus, if the integral
(9.4) makes sense, then our problems are solved. The proof goes as
follows:
CAN WE MAKE SENSE OUT OF "QUANTUM CHROMODYNAMICS"? 969
(9.8)
-
. (2.)
2. It&.f Q; =
JOO2CA) d. wW(CAJ)
2. [_two..
e -e i.c...>4l,]
o,z. 0
= _2. f oo e
_00
LWA
W.
CO W(w
~
..) ~
ClW
(9.9)
Thus, pea) and W(w 2 ) are each other's Fourier transform. The in-
verse of (9.9) lS
w(s) (9.10)
REFERENCES
16) D.J. Gross and F. Wilczek, Phys. Rev. Letters 30 (1973) 1343.
17) G. 't Hooft and M. Veltman, Nuclear Phys. B44 (1972) 189.
18) C.G. Bollini and J.J. Giambiagi, Phys. Letters 40B (1972) 566.
22) A.A. Belavin and A.A. Migdal, Gorky State University pre-
preprint (January 1974).
31) C.M. Bender and T.T. Wu, Phys. Rev. Letters 27 (1971) 461;
Phys. Rev. D7 (1972) 1620.
DISCUSSION No.1
BAULIEU:
't Hooft:
973
974 DISCUSSION
----~-------3i~
BUDNY:
't HOOFT:
COLEMAN:
There are exact soluble models having such parameters
and, in those cases, there is typically only one. There
may, of course, be any number but there is no reason
whatever to expect that there are more.
DISCUSSION 975
WIGHTMAN:
' t HOOFT:
COLEMAN:
PAFFUTI:
't HOOFT:
PREPARATA:
COLEMAN:
't HOOFT:
TOWNSEND:
't HOOFT:
WEEKS:
't HOOFT:
E
and in Minkowski space: ~ . t=.t
- I L..j
<Alt-')IA'(oJ) = L (AlC) IEi/<Ei \A'(t))le .
'E;
One should be the analytic continuation of the other.
COLEMAN:
PAAR:
't HOOFT:
the integral 1
Compare for instance the perturbation expansion for
DO d).
e~pl-~/A) -1+'2.
o
Here, it is not difficult to show thft the true value
always lies in between the nth and t 1e(n+l)th approxi-
I
WIGHTMAN:
't HOOFT:
't HOOfT:
LIPKIN:
't HOOfT:
LIPKIN:
't HOOFT:
LIPKIN:
't HOOFT:
WEILER:
't HOOfT:
COLEMAN:
't HOOfT:
BOHR:
It HOOFT:
TOWNSEND:
It HOOFT:
DO
A.S. Wightman
983
984 A. S. WIGHTMAN
a) it is internally consistent
b) there is a significant range of experiments which it
describes accurately.
This result does not exhaust the subject, but it shows how one can
give a precise meaning to the statement that the theory makes no
sense in the limit in which waves of arbitrarily high frequency are
allowed to interact.
(1)
"31-. 0 ...
o .
f ( x) a.n. ()(-4)
V\.
.
992 A. S. WIGHTMAN
~ {C X) =: 40 ~ f (a,+ )
tl~et
Thus the assertion that the gyromagnetic anomaly g-2/2 has the asymp-
totic series
n
,00
a (r:J./'if)n at 0 means no more and no less than that
Ln=
1
g-2/2 is defined for r:J. in some interval 0 < r:J. < r:J.o and has deriva-
tives of all orders from the right at O. Then the a are uniquely
n
determined; (2) holds for all N with f = g-2/2 and a = O.
for some c. a bound which one can derive from Cauchy's integral for-
mula for the a
n
exist. To see this, note that a state with a large number of elec-
trons of like charge could be made for which the binding energy
would be arbitrarily large and negative. growing, in fact, like the
number of electrons squared. If this electron cloud were combined
with a like cloud of opposite sign of charge and far away, one would
have a state of zero charge lower than the vacuum. Dyson then ar-
gued that the occurrence of such a vacuum instability would not per-
mit sensible Green's functions to exist, and therefore the hypothe-
sis that there is analyticity in the perturbation series at 0 is in-
admissible. There are at least two objections which can be raised
against Dyson's argument. First, it is not obvious that the analy-
tic continuation of the Green's functions to negative a are the
Green's functions of a theory in which like particles attract. In
fact, the Hamiltonian of such a theory would be of the form Ho iH I ,
non-Hermitian. Secondly, just because the vacuum is unstable it is
not obvious that the Green's functions fail to exist.
where
x
~
(>l-t:) Jt
On the theoretical side there are things that can be done if one
knows something about f. There are so-called summabi1ity methods.
These are operations on infinite series, divergent or convergent,
which yield convergent series or functions. Two elementary examples
are Cesaro summabi1ity and Abel summabi1ity. The first converts
it yields
Definition
The first important fact about this notion is that a given formal
power series has at most one function to which it is strongly asymp-
totic. Second, Borel's summability method always works for strongly
asymptotic series. More explicitly:
998 A. S. WIGHTMAN
Theorem
0\\ ~(Y\.)
(D+) ~
[ 2
fl::O (~!)2.
which converges for Izl < B and defines an analytic function F which
may be continued analytically to the sector larg zl < E. f can be
recovered from F by the integral formula
00
=
o
f .!..fi!!
t:+~
J tN\..df(tJ(Oo '\1. :: 0, f I 2.
Theorem
=
SHOULD WE BELIEVE IN QUANTUM FIELD THEORY? 1001
The second novelty about the theory was that it should exist
at all. Why should there be a general theory of quantized fields?
I think the answer ~s that the use of Green's functions on vacuum
expectation values to express the physical content of a theory (a
general idea for which J. Schwinger can claim as much credit as any-
body) is, on the one hand, very close to physical applications via
reduction formulas for S-matrix elements and, on the other hand, is
simply expressed in terms of the fields and physical vacuum. In
any case, it turned out that a rather impressive collection of gene-
ral results can be proved directly from the axioms without a detailed
use of the structure of equations of motion (PCT theorem, Spin-Sta-
tistics theorem, Haag-Ruelle scattering theory, some dispersion re-
lations, Goldstone theorem, etc.).
Here are some results of more than a decade of work by Glimm and
10)
Jaffe and their followers :
defined in L (-00, 00; dx) for -00 < a < 00, 0 < S < 00. The expansion
in which we are interested is a power series in S
(~) ~Y\..
E"o (0<,0+)_
I\<\.!
There was surprisingly little known about Eo(a,S) and this expansion
until the late 1960's when the following facts were established,
Eo(a,S) satisfies the scaling relation
- 2/'3 i\'\.
C"" (ol~ )
For A(~4)2 and AY2 the analogous questions have been studied
for the expansions in A of the Euclidean Green's functions. Notice
that these are the perturbation seri~es in the unrenormalized coupl-
11 12)
ing constant A. The authors' use a generalized version of the
above theorem on strongly asymptotic series which permits E =0 pro-
vided one has conditions on the boundary values of f(z) on the line
larg zl = rr/2.
these models have Schwinger functions which would blow up in the li-
mit in which ultra-violet cut-offs are removed if the renormalization
counterterms were not included in the Hamiltonian. Renormalization
is essential to make sense of these models.
c) The ingenuity and effort that went into the proofs of these re-
sults should not be underestimated. The proof of the existence of
the Euclidean Green's functions in A(~4)3 is the most difficult I
know of in mathematical physics. However, there ate also simple
general and enlightening ideas involved such as the method of cor-
relation inequalities.
(positivity)
and
w (1) = 1. (normalization)
it =
and observables leave each of the subspaces invariant. How are ob-
servations in one sector related to those in another? Local quan-
tum theory says that there is a quasi-local algebra of observables,
~, and that the representations of it in the various 4tne are unit-
ary inequivalent. Since the representations are also faithful (no
non-trivial observable is represented by 0) we can determine the
algebraic structure of ~ by looking at observations in anyone of
the sectors. What happens in the other sectors can then be con-
structed purely theoretically; one has just to construct the other
unitarily inequivalent representations of Gr. There is only one
proviso: not every representation is physically admissible. For ex-
ample, the representation may not lead to a unitary representation
of the Poincare group satisfying the spectral condition. A super-
selection sector is included in {t only if it is physically admis-
sible.
d
dimension
of
space-time
9renorm
cise if one compares the Euclidean expression for the vacuum energy
per unit volume in quantum field theory
Vi-'=> 60
with the grand canonical pressure 1n statistical mechanics.)
------~~~------+_--~x
and the singularity structure for a > 0 shown in Fig. 1. The exist-
ence of these singularities shows that the usual Borel summability
methods do not work on the anharmonic oscillator when the potential
has a double minimum. On the other hand, for the A(~4)2 model in
the two-phase region Glimm, Jaffe, and Spencer have established the
analyticity of the Schwinger functions and vacuum energy per unit
SHOULD WE BELIEVE IN QUANTUM FIELD THEORY? 1013
REFERENCES
10) For a general review, see the proceedings of the 1973 Erice
Mathematical Physics School, Constructive Quantum Field
Theory, Eds. G. Velo and A. Wightman, Lecture Notes in
Physics No. 25, Springer-Verlag, Berlin (1973).
DISCUSSION
DEO
You told us to believe in a physical theory if it is
internally consistent and has a fairly well defined area
of experiments for which it is accurate. I do not agree
that the second requirement is needed. Einstein once
said to the effect that the theory will predict what the
experiments will reveal. If theorists are guided solely
by experiments, we may have situations like the high
y-anomaly.
WIGHTMAN:
This is a comment rather than a question. I think
there are obvious reasons in favour of the second require-
ment as one wants to describe nature as it is.
DEO:
What is the predictive power of Axiomatic or
Constructive field theory?
WIGHTMAN:
There are general properties of QFT which have been
derived from first principles in the framework ofaxio-
matic field theory like the PCT theorem, connection be-
tween spin and statistics, dispersion relations, Goldstone
1017
1018 DISCUSSION
DEO:
WIGHTMAN:
People first chose the simplest models that they
could solve, and gradually complications were introduced.
Recently there has been a lot of work on the infrared
problem: first of all the work of Buchholz who showed
that there is a perfectly rigorous scattering theory for
massless particles. Also the work of Glimm and Jaffe on
the behaviour of QFT's at the critical point, where there
is no mass gap. There is a lot of work going on at the
moment about the infrared problem.
LIPKIN:
There is an interesting "interdisciplinary" corrunent
relating the lectures of Wightman, Gursey and Iachello
which also shows how the highly abstract work discussed
by Wightman has influenced bread and butter calculations
in particle physics. Both Arima and Iachello found SU(6)
syrrunetries in the real world of experimental data which
violated previous intuition. Iachello found a symmetry
in nuclear data suggesting a model of interacting bosons
DISCUSSION 1019
WIGHTMAN:
PRESKILL:
I am curious about the curve you sketched which gave
a zero of the ~ -function as a function of the dimension
in the (~~)Q model. I wonder how explicitly this
function is known. Do you just have a bound to the left
of which the curve must lie, or do you have more precise
information?
WIGHTMAN:
PRESKILL:
Do you have a heuristic understanding of the Glimm-
Jaffe bound?
DISCUSSION 1021
WIGHTMAN:
PRESKILL:
WIGHTMAN:
KENWAY:
WIGHTMAN:
KENWAY:
WIGHTMAN:
DINE:
WIGHTMAN:
CELMASTER:
WIGHTMAN:
TAO:
WIGHTMAN:
BACE:
WIGHTMAN:
WIGHTMAN:
M. Karowski
1. INTRODUCTION
1027
1028 M. KAROWSKI
o 4> +
-------------------------------
~===========---+_--~======~--~x
I'V\ == -f1~.S,,,,
in
The corresponding charges of asymptotic states !Pl, . , PN>out are
AN EXACT RELATIVISTIC S-MATRIX IN 1+1 DIMENSIONS 1029
i) no particle production;
ii) only momentum exchange, i.e., the sets of incoming and outgoing
momenta are equal;
iii) factorization of an N-particle S-matrix element into two-
particle ones:
(tJ)
S ( ?-1 I ~~
I )
where
1030 M. KAROWSKI
and
~l.
( -1 - 8'f'C )
h
were J~ = ;;,).l ,I,
0/ yo/an
dg '~s a coup l'~ng constant. The operator solu-
tion of this model is known since a long while and the Wightman
functions
In general, we have
= tl,.. F + iJ.\.c::..
1032 M. KAROWSKI
In momentum space on the mass shell after amputation only the parts
in ~ contribute which are linear in the fields. These vanish for
n
even n and are given for odd n by
+ _.
i) no particle production;
ii) only momentum exchange; and
iii) factorization of S.
0:- Jd 2 x <T~t1[~~Jtlc)() X)
12) ]..I
expects that the currents j can be redefined such that these
n
anomalies cancel. This was explicitly shown for n =3 13). We be-
lieve that the conservation laws and the properties (i), (ii), and
(iii) hold true in the quantized massive Thirring model.
3. THE S-MATRIX
Theorem
Proof
"'-1
~ (.~
~=i '~l ~W
The physical plane is mapped onto the strip 0 < 1m 6 < 1T and the
physical region for ff scattering (P1 + P2)2 - i > 4m2 corresponds
to 6 12 > O.
1036 M. KAROWSKI
itt
,---- - ---- - --- ------- - ---1
,
I I
I
T
. ...-.. . .
I
I C+I
=~---E===
o 4 m2 ~ L
I ___________ _ - - - - - - - ______1
o
r(9):. r(i.'ft-e)
yield
Hence, there remains only one unknown function tee) which we shall
now determine 1 ). If tn tee) is analytic in the physical strip and
does not grow terribly for IRe el + e, Cauchy's formula yields
-r
~ l(e) = -
1
--
AN EXACT RELATIVISTIC S-MATRIX IN 1+1 DIMENSIONS 1037
C being the contour enclosing the physical strip 0 < 1m 8 < TI. The
unitarity and crossing relations imply
'"
qe) = It)(P ~ J~)C
o
or
with
- ;( = 1 +
2.~
For A < 1, where there are no bound states, the S-matrix is com-
pletely determined. For A > 1 we have to calculate the scattering
of bound states, i.e., the amplitudes tbkf and tbkbi 2,16). We con-
sider the residue of S(3)(Pl,P2,P3) at (PI + P2)2 = ~ and obtain
after some calculations in agreement with previous results in the
' 1 aSS1ca
sem1C ' 1 I'1m1t
, 15) :
where
correspondingly for k ~ i:
where
4. OFF-SHELL QUANTITIES
17)
Using Watson's theorem , one can calculate the soliton form
IS)
factor
where
. :2.
4lM
>,.x..
--
n.
1:7
211:"""
o
ACKNOWLEDGEMENTS
REFERENCES
3) M.S. Ablowitz, O.J. Kaup, A.C. Newell and N. Segur, Phys. Rev.
Letters 31 (1973) 125.
L.A. Takhtadzhyan and L.D. Faddeev, Theor.Math.Phys. ~ (1975)
1046.
For a review, see: A.C. Scott, F.Y.F. Chu and D.W. McLaughlin,
Proc. IEEE ~ (1973) 1443.
12) P.P. Kulish and E.R. Nissimov, Pisma v. JETP 24 (1976) 247.
13) The first proof for this conservation law was done in loop
approximation:
B. Berg, M. Karowski and R.J. Thun, Phys. Letters 62B (1976)
63.
For a general treatment, see: R. Flume and S. Meyer, Nuovo
Cimento Letters 18 (1977) 236.
B. Berg, M. Karowskr-and R.J. Thun, Nuovo Cimento 38A (1977)
11.
15) V.E. Korepin and L.D. Faddeev, Theor. Math. Phys. 25 (1975)
1039.
J.L. Gervais and A. Jevicki, Nuclear Phys. B11D (1975) 113.
M.T. Jaekel, Nuclear Phys. Bl18 (1977) 508.
19) M. Gomes and J.R. Lowenstein, Nuclear Phys. B45 (1972) 252.
DYNAMICAL S~lliTRIES IN NUCLEAR PHYSICS
F. Iachello
1. INTRODUCTION
1043
1044 F.IACHELLQ
2. INTERACTING BOSONS
This situation has been changing in the last few years. The
idea is that one is actually not interested in all the levels of
the many-body system but only in a small fraction of them, those
having some specific structure. For instance, the nucleus 156 Gd
(2.1)
1046 F.IACHELLO
x )( X d L=2
x x x x x s L=O
(2.2)
3. DYNAMICAL SYMMETRIES
i) G == SU(S)
. ([ N] Y) d V 11 A L M) = EoJ -t 0( i nd (1'\ a- 1) +
(3.1)
f(nd-v)(nd+v+3)+~ [L(L+1)-6I\d]
ii) G - SU(3)
iii) G _ 0(6)
(3.3)
ii) (3.5).
iii) (3.6)
(3.7)
i) U(S-)
ex (2.) _ (elt
tt- S+5
t cd) (2)
f- -n (dtd)~}
2-
(3.9)
iii) O(~) (dfl);)) (d~!/)) (dls~$~)/2)
Examples of the first two kinds of symmetry are shown in Figs. 2 and
3, respectively. By comparing the experimental levels with those
provided by the corresponding classification scheme, one can see
that the symmetries are not badly broken. In fact, they appear to
1050 F.IACHELLO
1000
2+ 658 _2+ 722
1.5
4+ - 1298 - 1317 4 : - 1355 _ 1316
3+-1248 _ 1210
2+ - 1129- 1129 2-1154_1129
0+ - 1049 - 1049
1.0 8+ - 965 - 965
K=O K=2
(20,2)
6+ - 585 563
0.5
4+ - 288 268
2+ 89 80
0 0+ 0 0
K=O
(24,0)
E
(MeV)
1.5
1.0
0.5
82 86 90 94 98
Neutron Number
(4.1)
5. ODD-A NUCLEI.
GRADED LIE ALGEBRAS Al~D SUPERSYMMETRIES.
(5.1)
1054 F.IACHELLO
+
These operators are scalars in ordinary space but form a vector S
in a fictitious space, called quasi-spin. Similarly one can intro-
duce the two operators
(2. j -4- of ) 'I.,. t
V, :.
8 I/l.
~
tV
"'I
(5.2)
(2j + 1 (Ii
VI
"'~i:
= 8 '/2-
tt
'V
rL SA~ I Sb] == ~ _L
i C;~c. 5C a./b,c: 1,2,3 (5.3)
where S Sl iS 2 , S
Z
(5.4)
2.
6. CONCLUSIONS
REFERENCES
2) J.P. Elliott, Proc. Roy. Soc. (London) A245 (1958) 128 and 562.
F. Gursey
1. INTRODUCTION - An Overview
1059
1060 F.GORSEY
quantum numbers in a definite way accounted well for the low lying
hadron spectrum 3 ). This could be understood if the hadrons were
assumed to be some sort of composite structures of quarks, 3 quarks
for a baryon and a quark-anti quark pair for a meson 4 ). However
the mechanism for such a binding remained completely mysterious.
The crucial quark concept turned out to be very useful in weak and
electromagnetic interactions of hadrons since many of their
properties could be better understood if it was assumed that hadrons
SYMMETRIES OF QUARKS AND LEPTONS 1061
1. g~~~!~~!!~~_~!_!~!~!~~!_~~~_~E~~:!!~~_~~~~!!!~~
2. ~~~E~E~~l_~~E~E~EiY~_~~_Y~~!Ei~~_~i~~~g_Ei~_~~_l~!~E~~l
g~~~E~!!.l_~~~E~
(2.2.1)
1070 F. GURSEY
(2.2.2)
(2.2.3)
namely, the wand p vector mesons and the pions that form the
15 dimensional multiplet of the adjoint representation of SU(4).
The resulting spin independent potential was applied to N-N scatter-
ing and the deuteron by these authors who were unaware of Wigner's
work and failed to recognize the group structure of their theory.
Again, without the use of group theoretic language, Rarita and
Schwinger introduced a symmetry breaking in the theory by keeping
the vertex symmetry of the couplings to obtain a well-behaved
potential for the Schrodinger equation but by breaking the symmetry
through a mass difference in the 7T and p mesons. They were thus
able to explain the electric quadrupole moment of the deuteron
49) .
which vanishes in the exact SU(4) symmetry limit. Jauch , us ~ng
the same theory showed that the best fit to experiment is obtained
if the p meson is a few times heavier than the pion.
After the wand p mesons were well established, and the SU(3)
flavor symmetry gained acceptance through the completion of the
pseudoscalar and vector octets, it was time to redo the work of
the early forties on the baryon-baryon potential by allowing them
to exchange vector and pseudoscalar octets, replacing isospin by
SU(3) and upgrading Wigner's SU(4) symmetry to SU(6) with the
35 generators
The corrections to the free quark SU(6) come chiefly from the
distortion of free particle wave functions due to the confining
potential. These effects can be calculated in the bag approx-
. .
~mat~on
58,59) and resu It~n b et t er agreement w~t
. h exper~ment.
.
3. ~~~~~~~~!~gi~!_Y~!i~i~~_~!_~~_~EE!~!i~~~~_~gi2_~_~Qi~2orbital
:_Q~_~~_g~~~!2~~~~_!~i2_~~~!ll
. 67)
a) SU(6) x SO(3) symmetry of the mult1plet structure
qq meson multiplets that are experimentally well established:
l(L=O); 35(L=0); 35(L=l).
qqq baryon multiplets that are experimentally well established:
56(L=0); 70(L=l); 56(L=2).
There is also the new possibility of the existence of
qqqq exotic mesons that would fit into 405(L=0) or 189 (L:O)
Examples are:
The ratio of the proton and neutron magnetic moments ~p/~N is
-3/2 from SU(6), the experimental value being -1.46. Related
results are the vanishing of the charge form factor of the neutron
and the constancy of the magnetic and electric form factor ratio
GM(t)/GE(t) where t is the invariant momentum transfer.
For the gA/gV ratio the exact symmetry limit gives 5/3 which
is too high. However confinement corrections bring it down to
59
values between 1.20 and 1.30 in agreement with experiment ).
From this brief survey we can conclude that the SU(6) structure
is undoubtedly present in hadron physics. The question is:
how far is it understood and how fundamental is it?
i
1l11J!, (i 1,2,3), (2.4.1)
where m is the quark mass, i the color index and D~ the covariant
derivative.
~ a a
D ' \ + zgA BI-l' (a 1, . ,8), (2.4.2)
).l
(2.4.3)
where
A
b :: (2.4.4)
(2.4.5)
if Y5 ~s diagonal, (2.4.6)
and
if Y4 ~s diagonal. (2.4.7)
1m, s
-7
,k.
2'
p, S >
Z (2.4.8)
with
l
m (2.4.9)
SYMMETRIES OF QUARKS AND LEPTONS 1081
(2.4.10)
-+--l'.!
{ x;:; r(.)} -t+~lf.
ex.p "2 u~~) f\ P
I
-2-
_ ,/PO(tt1+i.~'e)21+~1t
- V~ f.. 2' (2.4.11)
The projection operators 1/2 (1~4) project the large (+) or small
(-) components of the free wave function. To simplify the formulae
we have also omitted the color index.
a) Quark Number
(2.4.12)
where
.
J'~ ;:::
-L\((Jl
Ip U r ) Col = 1)2,3 ) . (2.4.13)
J p )
b) Charges
(2.4.14)
1082 F. GORSEY
c) Energy Momentum
(2.4.15)
(2.4.16)
(2.4.17)
(ia
n
= pn on a momentum eigenstate).
(2.4.18)
and the identity (2.4.11). To obtain the covariant state from the
Wigner state we must then apply a non-local operator on
I p > exp(i p.x). We find
'P'lo>:= W(-i. :~1).f ~fe~P'\4~p)ID (2.4.20)
SYMMETRIES OF QUARKS AND LEPTONS 1083
where
W(-I,. V)
........
(2.4.21)
(2.4.22)
c
creates Wigner antiquark states and ~ (x) creates quark states.
The non-local operator W(-iV) is a unitary operator that connects
covariant states with Wigner states.
(2.4.24a)
or
(2.4.24b)
(2.4.25)
(2.4.26)
1084 F. GURSEY
It l.S given by
W := (2.4.27)
J.
where
..... -
~.L' p~ .: ~, P. +~.! F:2. '
./2.
F'.l. = V P. + P:4 ,
2
(2.4.28)
f" .= - i 6n (2.4.29)
Note that HI! no longer commutes with n and Y40 n but only
with the subset
(2.4.31)
(2.4.34)
where
- -
Ht' yA = -t ~ ,
'f cr')A\> ()~ ~ = -~ E)L))A~ If 6, ~f '1>+ add. terms
(2.4.35)
is conserved:
so that the n~v are time independent. These are related to the
Wigner spin as
-I
'vJ SG OL W = 0 ~
0
-2IE ..
lJ K
W-I .Jc.
n. u W = 52:.:.-
J '" l 2.
'y(t a.o.
a.. P L P
-to bt d. b)o3
r L
dp
p (2.4.37)
1086 F. GURSEY
Hence they can be interpreted as quark spin plus anti quark spin and
are associated wi th the d ~
1
(2.4.38)
where
and
(2.4.40)
Jl:::. W-'n
t.
5i W ~ ~f( ~t cr. a. - bt G, b) io
pl.,. pl.~ I'
(2.4.41)
-1
W 52!.;o W-O. (2.4.42)
Sl:,JA-' the C-odd operators .l2.i. Sl.; and the C-even ~ COIImlute wi th the
Hami 1 tonian H".
and the C-even subgroup of SU(6) becomes the chiral SU(3) x SU(3)
group whose generators COIImlute with massless Dirac Hamiltonian.
so that
li tn [SLJ(3)
~~ "~DO
, (2.5.1)
(2.5.2)
(2.5.3)
where.t (lji), (p) and.,t' (WlI) are respectively free Lagrangians for
o 0 0 ~
(2.5.4)
(2.5.6)
+ total divergence
(2.5.7)
The first term is of the order ).l 12M if).l is the meson mass and
negligible in the static limit. In the second term the operator
Y4 Y5-+
Y -+
-ocommutes
= ' h Y an d as a resu ltoes
W1t d not m1X. part1c' 1e
4
states with antiparticle states, so that putting 1;+ = I; (particle
wave function) we can make the replacement
(2.5.9)
, (2.5.11)
(2.5.12)
(2.5.13)
(2.5.14)
(2.5.15)
(2.5.16)
(2.5.17)
(2.5.18)
with
(2.5.19)
-a
all and 0.L generate SU(2\ and combine with an internal SU(N) to
give SU(2N) Performing a Cini-Touschek-Melosh transformation in
-r w
the k direction for the kinetic term we see that the generators of
this SU(2N) commute with the transformed Hamiltonian including the
w
effective approximate interaction term.
(2.5.20)
(2.5.21)
SYMMETRIES OF QUARKS AND LEPTONS 1093
(2.5.22)
The exchange between quarks will give for the Fourier trans-
-+
form v(k) of the potential
(2.6.1)
-+
where k lS the space-like momentum transfer (in a frame with k
o
= 0),
m lS the meson mass, M the quark mass (that becomes the baryon mass
when we add the masses of the spectator quarks) and f the pseudo-
scalar coupling constant. Its Fourier transform gives the potential
73)
due to , namely
(2.6.2)
x= mr (2.6.3)
3 C'(I)
6 r ....)((5 .+)
.. (2)r
5 i2- ;: r2.
...., (,) - (2.)
(5 cr , (2.6.4)
with
- I'~ I,
.~
r r -
I
('
2 (2.6.5)
and
. -J(
(2.6.7)
(2.6.8)
(2.6.9)
(2.6.10)
(2.6.11)
(2.6.13)
SYMMETRIES OF QUARKS AND LEPTONS 1097
~Il (2.6.14)
(2.6.15)
that commutes with the SU(4) generators found from ~(~(ll +~~)) and
~(~(~ +~(~). This potential is therefore spin and isospin independ-
ent in the limit of degenerate masses and equal coupling.
. ~ i
If the mass degeneracy is lifted by al 1 ow~ng 'IT and p to have
different masses, a tensor force S12 occurs in the potential between
u and d, hence, between the proton and the neutron, so that the
deuteron acquires an electric quadrupole moment. The right sign
an d .
Qagn~tu d e are found 49 ) ~f t h e P ~s
~ . tree
h to f our t'~mes h eav~er
.
than the pion, in accordance with experiment.
-
L-
.-+ -+
r x. I:>
-
5 .- (2.6.16)
1098 F. GORSEY
(2.6.17)
(2.6.18)
++
The exchange of the pseudoscalar ~ or Q12
contributes no L.S
terms that may cancel the highly singular additional terms due to
w exchange. A Q12 term that can cancel the Q12 term can only come
from the exchange of a scalar meson a. Accordingly we can enlarge
our scheme by adding a scalar field a with the Yukawa interaction
(2.6.19)
(2.6.20)
SYMMETRIES OF QUARKS AND LEPTONS 1099
V= v' + VW + yO'
the Q12 terms cancel, leaving
(2.6.21)
-
generators of SU(4) are
w
Note that the cancellation the Q12 terms requires the degen-
~f
l -L
T ....... Il ) J
(2.8.1)
which give the usual SU(4) classification for the bound states qq.
The spin dependence of the one-gluon exchange between q and q
. -+(1) -+(2)
~ntroduces a a .a term that splits TI from wand p and leads to
the mass formula for SU(6)
1102 F. GURSEY
that was proposed 3 ) in 1964. The M part comes from the exact
o
5U(6) limit valid for the spin independent confining potentia1 7S ).
This spin independence is only seen in the lattice approximation in
QCD. We have found a partial justification in the Moeller-Rosenfeld
model with quarks exchanging almost degenerate qq bound states.
The 5(5+1) part comes from short range one-gluon exchange forces
as shown by de Rujula, Georgi and Glashow 3S ) who find the value of
M3 to be proportional to as' the QCD fine structure constant and
inversely proportional to the masses of the constituent quarks.
Finally the Ml and M2 terms come from quark mass difference.
The quark fields are assumed to obey field equations and inter-
act locally with boson fields. These fundamental bosonic fields
are of two kinds. The ones associated with the exact color group
are obtained as gauge fields arising from the local validity of
SU(3)c. They are 8 massless vector gluon fields in the adjoint
representation of SU(3)c which only interact with quarks and with
each other and consequently are responsible for the strong inter-
actions of quarks. Quarks also interact with a massless photon, a
vector gauge field associated with a U(l) group. But the photon,
unlike the gluons, does not interact exclusively with the quarks.
The same is true for the massive weak bosons (yet to be discovered)
that are responsible for the weak interactions of quarks. There
+
are at leas't 3 weak bosons. Their associated vector fields W-, Z ,
11 11
together with the electromagnetic field potential A can be inter-
11
preted as the gauge fields of the Weinberg-Salam group SU(2) x U(l)
that combines weak and electromagnetic interactions. Quarks are
also assumed to interact with scalar fields that belong to some
representation of the flavor group but are singlets under the color
group. These Higgs scalars break the flavor group spontaneously,
keeping only the electromagnetic U(l) exact. This process of
spontaneous symmetry breaking through the vacuum expectation values
of the Higgs fields gives masses to some flavor gauge bosons, for
example to the three weak bosons. It is not clear whether the
Higgs fields are fundamental or are just effective fields arising
from qq bound states. If the local gauge symmetry is extended to
local supersymmetry, then the existence of fundamental Higgs fields
1106 F. GURSEY
(3.1.1)
(3.1.2)
, ( :~) , (3.1.3)
(3.1.4)
sL'
SYMMETRIES OF QUARKS AND LEPTONS 1107
(~:) ,
I-
(;:) (;~) , (3.1.5)
(e~) . (3.1.6)
Again, the definite assignments for ).IR and 'R are not known. In the
simple version of the W-S theory one assigns all the right-handed
leptons to SU(2) singlets so that we can complete (3.1.6) as
(3.1.7)
Suppose the assignments of the simple W-S scheme are not cor-
rect but (3.1.1), (3.1.2), (3.1.5) and (3.1.6) still hold. We
1108 F. GURSEY
(3.1.8)
AI
dR 0
A"l> 1\2
Utt -lA./\ -d.' I..
I
iLl..
,,3
-Ll.jl 0
AI (\1 \.lI-
Ce)
a~ (e)
/..lit - l-
dR3 ,
"Zo AI
0 -d~ 143
"e
l-
d.'I.. d2.L d1I. 0 "
eR.
-u.' _v..1 _tiJ _~ 0
e~ 1-1- I. A. (3.1.9)
SYMMETRIES OF QUARKS AND LEPTONS 1109
e
of the group G. The quark indices refer to color. The SU(2)weak
generators are
l ') 0 )
,-, 1(')_-- ( 0I 0
J
0)~ .
o '2 T 0 0 (3. 1. lla)
C-A 0)
~ lil'
I (2.)=.
(~ 0)
I ' (3.1.11b)
where Aa are the 3x3 Gell-Mann matrices for SU(3), and the electric
charge operator in SU(2) x U(l) is
Q=
3
0
0
0
I
0
0
0
I
3
0
0
I
:3
a a
0 0
0 0
()
0
==
(i
0
1(') 0 )
-4 (,-T3)
0 (3.1.12)
0 0 0 0 -I
Sln
. :2.6 w
(3.1.13)
are associated with new vector bosons in a gauge theory. These new
bosons acting on ~(5)
e into -dR'
will transform e L and v L Since they
carry both lepton and quark quantum numbers, they are called lepto-
quarks. If we consider the action on ~(10) we see that some bosons
1110 F. GURSEY
will change quarks into antiquarks. Thus they will behave like
diquarks or antidiquarks. Some bosons will be simultaneously
diquarks and leptoquarks. When such bosons are emitted a quark can
turn into a lepton or an antiquark. It follows that by the exchange
of a diquark-leptoquark boson, two quarks can turn into a lepton
and an anti quark.
(3.1.15)
There remains one other question. The Planck mass is the mass
that comes in Quantum gravity. Hence, a theory involving such a
large mass cannot ignore gravitation. The large mass scale may make
sense ~n a superunified theory of strong~ electromagnetic, weak and
1112 F.GURSEY
The smallest groups in which SU(5) can be embedded are SO(lO) and
SU(6), both of rank 5. The SU(6) assignments have been recently
studied by Segre et al;4). In order to cancel Adler-Jackiw-Bell
anomalies, two different representations of SU(6) are required,
namely two six-dimensional and one IS-dimensional representations.
These would correspond to the
o 1\ 2 .'
-u R. -dL IA,. bl-
I I
I
I
I
.(! I
r::;.1- /
(3.1.19)
(~) (~) ~ M
d, b) replaced by (a ,S ,v, v , ~, M, c, s, h). Thus, the
SU(6) generalization does not require a charge 2/3 top quark, but
a new charge -1/3 heavy quark h. The embedding of SUeS) is shown
in dotted lines, so that the SU(2)weak is embedded in SU(6) via
1114 F. GURSEY
::
from (3.1.19). The doublets are
with all the rest being singlets. Similarly, for the muon family:
(3.1.21)
(3.2.2)
with the conjugate decomposition for the 77. Thus the 27 describes
9 leptons, 3 colored quarks and 3 colored antiquarks. Two (27)
describe 18 two-component leptons with the charged sector corres-
ponding to e, T, ~ and M, and 6 colored quarks with charges corres-
ponding to u, d, b, c, sand h. To see how the W-S group is embed-
ded in E6 we consider the decompositions b, c and d
Thus, we find automatically two (6) and one (15) of SU(6) and
(TO) and (5) representations of SU(5) with additional (5) and (1)
representations. The SO(lO) decomposition includes the spinor re-
presentation 16, but there are also additional (10) and singlet
representations. It follows that the SU(6) assignment (3.1.19) ~s
also valid for the (27) of E6 The SU(5) assignments can also be
read off after dividing the representations by the dotted lines as
shown. The weak SU(2) doublet structure is therefore given by
(3.1.20) and, for the muon family by (3.1.21). This assignment is
. 1e
compat~b . h
w~t .
exper~ments so f ar.87) Th e d'~scovery 0 f '
par~ty
(3.2.8)
1118 F.GURSEY
The first two columns are the weak leptons doublets. The (3,1)
quarks and the (1,3) antiquarks are
(3.2.9)
with the first two lines of the quark triplet being in weak doublets.
Thus the weak SU(2) is in the first flavor SU(3) group.
Similarly for the muon family we have the SU(3) x SU(3) decom-
positions
(3.2.10)
(3.2.11)
The SU(2) singlets are (35) and (15) of SU(6). These are just
the Higgs multiplets considered by Segre and Weldon.
(4.1. 6)
(4.1.8)
We have
(4.1.12)
It follows that
U( Q) p~ =- {Q~Q} (4.1.13)
J
(
= ~ (4.1.15)
>
(4.1.19)
1S the determinant of J.
Matrices PEJ 8 which obey the conditions
3
? X p:= 0 (4.1.26)
and
Tr P =1 (4.1. 27)
can be shown to satisfy also
) (4.1. 28)
so that they are projection operators. They have the form
(4.1. 30)
(4.1.31)
(4.1.32)
(4.1. 33)
-.
X,:=, 0<,CX 3 , , (4.1. 34)
(4.1.35)
that obeys
(4.1. 36)
I=- 2:
L
IL><~I ==:. ~
L
P:
L
, (4.1.38)
so that the relation
(4.1. 40)
[AJ BJ =-4[(A,BJ,J] J
(4.1.41)
[A J J3 J = 0 (4.1.42)
1130 F. GURSEY
implies
2. ~~~~E!~~~~_~E~E~E!i~~_~!_!~~_~~~~E!i~~~1_9~~~!~~_~~~~~~i~~-~!42~
~~~~!~1i~~!i~~_E~_9~~E1~~_~~!~~~_~1g~~!~~_~~62~
(4.2.1)
when
t
(IX'>= U/oc> , 1/3 1
) = UI~>, UU =t (4.2.2)
Q1::UQu t , (4.2.4)
(4.2.6)
(4.2.7)
Exponentiation gives
J': J + [i.H, J] +r!L~H,(iH,JJJ+ ,.,
= Cexp~H)J ex.p(-iH), (4.Z.9)
which shows that J is transformed by a unitary matrix. In the
octonionic case the finite transformation of F4 is given by the
series
(4.2.10)
which only involves the Jordan product and can no longer be written
in the form (4.2.9). When HI involves only one octonion and H2 is
a purely scalar matrix then (4.2.10) can be integrated in the form
(4.Z.9) with iH replaced by an antihermitian octonionic matrix
involving one octonion only. It is seen that the full group is
1132 F. GURSEY
r1 -::: Tr J ) (4.2.11)
12 :::. Tr J2 , (4.2.12)
" iI
..... 0 ...0
It = 0, (4.2.16)
(4.2.17)
(4.2.18)
(4.2.19)
together with the complex conjugate equations. Here one sees the
virtue of octonion multiplication. If we consider the direct
products
'3 x 3:::.3+6, (4.2.20)
'3X3 t+~ (4.2.21)
c
for SU(3) , then these equations show that octonion multiplication
gets rid of (6) in 3 x 3, while it gets rid of (8) in 3 x 3.
Combining (4.2.18) and (4.2.19) we find
Thus, the octonion product leaves only the color singlet part
in 3 x 3 and 3 x 3 x 3, so that it is a natural algebra for colored
quarks.
(4.2.24)
decomposition
(4.2.25)
-F T= U*LT+
o U.
0
L- u~QJ:r-u.f(~T=
J J J
F (4.2.26)
(4.2.27)
(4.2.28)
(4.2.29)
(4.2.31)
14 = (F )( F, F* X F'It'). (4.2.33)
A geometry which generalizes the projective geometry of the
Moufang plane can be based on the complex matrices F. It is called
.. 1 102) .
t h e geometry 0 f comp 1ex octon10n1C panes. A genera11zed p01nt
(or state) is defined by S such that
SxS o (4.2.34)
and is E6 invariant.
Acknowledgments
REFERENCES
51. S. Weinberg, Phys. Rev. 133B, 1318 (1964), Phys. Rev. 139B,
597 (1965).
52. L.O'Raifeartaigh, Phys. Rev. 139B, 1052 (1965).
53. B. Stech, Zeit. Phys. 144, 214 (1956).
54. H.J. Lipkin and S. Meshkov. Phys. Rev. Lett. ~, 670 (1965).
55. R.P. Feynman, M. Ge11-Mann and G. Zweig, Phys. Rev. Lett. 13,
678 (1964);
R.F. Dashen and M. Ge11-Mann, Phys. Lett. 12, 142, 145 (1965).
56. For a review, see K.G. Wilson, in New Pathways in High Energy
Physics II, ed. A. Perlmutter, p. 243 (Plenum, New York, 1976).
57. J. Me1osh, Phys. Rev. D9, 1095 (1974).
58. P.N. Bogo1iubov, Ann. Institut H. Poincare ~, 163 (1968).
59. A. Chodos, R.L. Jaffe, K. Johnson and V.F. Weisskopf,
Phys. Rev. D9, 3471 (1974).
60. L. Fo1dy and S.A. Wouthuysen, Phys. Rev. ~ 29 (1950).
61. F. Gursey, Phys. Lett. 14, 330 (1965).
62. F.J. Gilman, M. Kugler and S. Meshkov, Phys. Rev. D9,
715 (1974).
63. A.J.C. Hey, J.L. Rosner and J. Weyers, Nuc1. Phys. B61,
205 (1973).
64. F.E. Close, H. Osborn and A.M. Thompson, Nuc1. Phys. B77,
281 (1974).
65. L.H. Chan, Phys. Rev. ~, 2478 (1977).
66. R. Friedberg and T.D. Lee, Phys. Rev. D16, 1096 (1977),
also: "QCD and the Soliton Model of Hadrons" Columbia
preprint CU-TP-118 (1978).
67. J. Rosner, Phys. Reports 11C, 190 (1974);
O.W. Greenberg and C.A. Nelson, Phys. Reports 32C, 69 (1977).
68. J.K. Kim, Phys. Rev. Lett. 1I. 1079 (1967);
C.H. Chan and F.T. Meiere, Phys. Rev. Lett. 11, 299 (1964).
1144 F.GURSEY
DISCUSSION No.1
MOTTOLA:
GORSEY:
1147
1148 DISCUSSION
where
JANCEWICZ:
GURSEY:
J T-+
ap cr-Qp
_13
01 1> (and
5b+p -+ Dr d t"'
Consider a single Dirac particle.
(j'
~
The spin operators
for antiparticles) are
conserved so that they can be considered as internal
degrees of freedom which generate an SU(2) algebra. These
operators !~e not ~ovariant however. The covariant oper-
ators are '\l ~... '\> d x but these are not conserved since
they do no commute with the Dirac Hamiltonian :
II ... -
n="'4m+ l.'("{.V. Under a FOldy-Wouthuysen transformat~on
H _ '1C = lH,,~\)-1 and the covariant spin is replaced by the
non-covariant spins. The non-covariant spins do commute
wi th 1< so they are conserved.
If we have two Dirac particles in a doublet of SU(2)
isospin then the following 15 non-covariant operators
are conserved
JQ+pOi.Q p d p, SQ.tp~iap d'3? j a+r <!i1;j4.\' d'jp
3 I
TOWNSEND:
GORSEY:
CARTER:
For a Dirac spinor there are two separately conserv-
ed spins for particles and antiparticles because a Dirac
spinor is not irreducible under the Poincare group. But
for an interacting Dirac particle these spins are not
separately conserved. Is this related to the need to
write particle and antiparticle in an interacting field
theory and what is the status of interacting fields with
respect to the Lorentz and Poincare group?
GORSEY:
The representation theory of the Poincare group as
derived by Wigner is purely for free particles. Starting
from these representations and multiplying them by cer-
tain non-unitary matrices you get the covariant fields.
"Covariant" means "finite dimensional representation of
the Lorentz group" ; the Poincare group has nothing to
do with covariance, its unitary representations are non-
covariant as we have seen. In field theory we also want
the fields to belong to finite, non-unitary and linear
representations of the Lorentz group and this is a diff-
erent requirement.
We impose this condition because we want locality,
necessary for microcausality. The Poincare group does
not act locally on particle states but acts locally on
the fields. This is why, when we reduce a field with
respect to the Poincare group we get all sorts of part-
icles. For example a vector field describes two part-
icles, a 1- and a 0+. We suppress the 0+ with the con-
dition ")",A", = o. For a spinor field we get both part-
icles and anti-particles from the same field.
It is interesting that for the Poincare group Wigner
showed that you can find 3 operators Xi functions of
the Poincare generators, which have the remarkable
1152 DISCUSSION
property that
MIETTINEN:
/
quarks
-----
----
heavy
medium heavy
light
'"
fermion
----
heavy
leptons ~ medium heavy
light
<
/
quarks
fermion med~um heavy
\
leptons
heavy
------ quarks
leptons
GORSEY:
It is true that in the context of gauge theories,
the IIsplittingll occurs in two stages which correspond to
the first diagram you have drawn. But this initial
splitting has nothing to do with fermion masses. In the
first stage, we break the symmetry and allow the gauge
bosons to acquire mass. Now it is an experimental fact
that quarks do not transform themselves into leptons
at observable rates. We therefore require that the
gauge boson mass splitting occurs in such a way that the
bosons which transform quarks into leptons be very heavy.
So we have indeed separated quarks and leptons in the
first step, but this separation, is forced upon us by
the stability of the proton and occurs by introducing
asymmetry in the gauge boson sector, not the fermion
sector. In a sense, the symmetry of quarks and leptons
is preserved: they have very similar interactions. They
just do not mix with one another. Also, as I showed, the
quark-lepton symmetry is not necessarily expressed in a
doublet series as you wrote it.
MIETTINEN:
Today we usually leave out gravitation when we begin
to try to unite the interactions. Why do we think it is
better to begin with strong, electromagnetic, and weak
interactions, and save gravity for the future? Einstein
began with gravitation.
GORSEY:
In Einstein's time, only two forces were known:
gravity and electromagnetism. Therefore unification nec-
essarily involved gravitation. Another point concerns
Einstein's philosophy. He wished to do away with the
concept of a force and reduce all dynamics to geometry.
Maybe if Einstein had known about the weak and strong
forces, he would never have approached unification this
way.
KENWAY:
I would like to ask you to comment on the justifi-
cation for considering the unification of non-gravitation-
al interactions within an exceptional group structure.
DISCUSSION 1155
GORSEY:
~ICHICHI:
GORSEY:
RAJPOOT:
If you do allow a particle to have both right-handed
and left-handed fields, then it is not clear that you can
keep it massless. Even if in the bare fermion mass mat-
rix the field has zero eigenvalue, it can acquire mass
in higher orders through radiative corrections.
GORSEY:
BUDNY:
GORSEY:
\
1+ ~ ex. )
($iY'l2~W ')
otsno~ma"2~c.I
. =A I (
BUNK:
GORSEY:
If we use the Higgs 912-plet there are only two
parameters because one can only construct a single quad-
ratic invariant potential from the 912.
BUNK:
GORSEY:
The two parameters correspond to the mass levels of
the vector bosons: they are used to induce this splitt-
ing. Both are very heavy. Their only effects at low
energy are in the renormalization effects that fix ~s
and SiY)z~w.
RAJPOOT:
What about the left-OVer Goldstone bosons in your
model?
1158 DISCUSSION
GORSEY:
RAJPOOT:
19
What is the meaning of the Planck mass, 10 GeV,
which arises in unified gauge theories of strong, weak
and electromagnetic interactions?
GORSEY:
BACE:
GORSEY:
CELMASTER:
G'ORSEY:
If you do not put quarks and leptons into the low-
est representation, you are back to the old problem of the
Eightfold Way. If we put the quarks and leptons into a
larger representation we would be confronted with the
question of what the lowest representation meant. Higher
dimensional representations can always be built from the
lowest-dimensional representation. In the Eightfold Way,
one puts hadrons into higher-dimensional representations
of SU(3) (i.e. octet, decuplet), and the concept of
quarks emerged from wondering what the triplet represent-
ation meant. So you see that if we put quarks and lep-
tons into such a representation, we would have to wonder
whether they were constructed out of objects which be-
longed to the lowest-dimensional representation! The
whole idea of whether quarks and leptons were the funda-
mental fields would be in doubt.
BIGI:
Do you put both fermions and antifermions in the
lowest dimensional representation?
GtJRSEY:
Yes.
BIGI:
Then do you not have fermion number non-conservation
in addition to baryon number non-conservation?
G'ORSEY:
Not at all. That is why I insisted that the gauge
group be complex, so I would have the invariance under
phase transformations with which we characterize ferm-
ionic multiplets. Thus a fermion will not decay into a
boson. Nevertheless, just like baryon number, lepton
number will also be violated in general since some of
the generators of the group that connect leptons to anti-
leptons will have lepton number 2. The associated bosons
1160 DISCUSSION
CARTER:
GORSEY:
LIPKIN:
GORSEY:
This is an illustration of the fact that SU(6) is
not an invariance of the S-matrix.
GORSEY:
HINCHLIFFE:
Can you comment on the status of supergravity and in
particular on the status of exceptional groups in attempts
to unify gravity with other interactions?
GORSEY:
I am not an expert on supergravity or supersymmetry.
What I know is what a layman knows, among theorists. All
the simple Lie superalgebras have been classified by Kac,
Kaplansky, and Freund by assuming two axioms : 1) You
have anticommutators as well as commutators, so you have
a generalized Lie algebra, a graded Lie algebra. 2) You
have something which corresponds to the Jacobi identity,
and this is obtained by taking the usual Jacobi identity
and putting in some anticommuting numbers, assuming that
these anticommuting numbers are associative. Then you
get a generalized form of the Jacobi identity which will
hold for generators which are anticommuting. Once you
have these two axioms you can deduce all systems which
satisfy them, and these have been classified. You also
want simplicity, so you look for the simple systems.
You generalize the Killing metric, and demand that it be
non-singular. This then gives you all the simple algebras.
1162 DISCUSSION
CARTER:
You have said that if one investigates supersymmetry
algebras derived from generalized Jacobi identities which
are in turn derived using non-associative algebras, i.e.,
octon-ions, then you might find that these supersymmetry
algebras have exceptional subalgebras. Then we could
unite gravity with an elementary particle theory based on
an exceptional group. That would be nice, but would the
fact that you used octonions to derive the generalized
Jacobi identities mean that you had to do all the physics
with octonions?
1164 DISCUSSION
Gt'JRSEY:
MOTTOLA:
GURSEY:
MIETTINEN:
GORSEY:
A. Sanda
Fact 1
'1/~
r ? t> p ?
1165
1166 A. SANDA
(c) leads to
Let us take (a) as our model for J production. Then extending the
simple model,
:rN ~: q
can be related to
We obtain
('=
THE BEST WHY 1167
and
First get
q. 'Z
-
~I~I =
21~r
\~:r I~r
Using this ratio
1168 A. SANDA
('"-
2M~ d::M Z (1)~-+e"'e--+x) 2m-;r
~
(~~)
2.
;
=
~ S eN-
ee d~dM2
dM~(p~~J+)() _1 S
See
C~)~
of; ltm.,-~ 1
-~ee
.f'a
-L = 11
.of1\"
we thus get
r= - j
THE BEST WHY 1169
Fact 2
~e"h -- 1 ;Co 1
".... 200
important? (No, because this would give (pp pp) for J production.)
Do we understand the IZO rule?
STATUS OF THE SUBNUCLEAR WHYs
Antonino zichichi
CERN
Geneva, Switzerland
INTRODUCTION
1. Do We Understand Isospin?
2. Do We Understand Strangeness?
1171
A. ZICHICHI
1172
_d
rLcms
if)
20
f-
+r---t
Z
w
>
w
0
w 10
f-
I
t++t
<.:)
w
3:
0
-I 0
COS EJ *
Fig. 2.1 Experimental results on the Q-
spin measurement from the ABCLV Collaboration
STATUS OF THE SUBNUCLEAR WHYS 1173
+ - .
p + nucleus + ~ ~ + anythlng
has produced a new bump, called T, at H(~+~-) '" 9.5 GeV with
FWHU '" (1.16 0.09) GeV and a cross-section
Conclusion: (J/l}J) and continuum scale in the same way. Only the
"coupling" strength differs by a factor of '" 200. Notice, however,
that to the J/l}J inclusive production there contribute other particles
decaying into (J/l}J). So the factor (200) is certainly the upper
limit for the "coupling" which produces (JI1/!) with respect to the
continuum. Repeating the same comparison for l}J' production we get:
[M 2 (dO/dy)]JI1/!
'" 6
[M2(dO/dy)]l}J1
10 31
t I [aUncertainty
C0 bb eo. 66"1. scal1
o Snyder et 01.
D Bronson et 01.
1032 A Busser et 01 .
Nagy et 01.
... Antipov et 01.
x Aubert et 01.
N
E 16 33
0
0
">- "
"t"t, IjJ'
bl>-
"0 "0
CD 1634
'y
/\
16 35
B dCTI =7.511034el47MV\ ,
dy y=o
1036
a .1 .2 .7
Fig. 3.1 Dependence on beam energy: B da/dy for J/~ and ~'
production in hadron collisions
1176 A. ZICHICHI
4.2
",(4100)
-------- ..... CHARMED HADRONS
4.0
THRESHOLD FOR
\ OPEN CHARM
3.8 \ 3.73 GeV '" (3772) n 3 01
._.- .. ----_._-_._._-_.-._ .. -.--_._._._._._-_._.
'" (3685)
>
~ 3.6 ____Lo_____
'P? X (3555)
CI)
X (3510)
CI)
X(3455)
~
2 1S 0 1 \ X (3415)
1 \
~ 3.4 I \
\
,
I
1 \Y 11'11'
t
HADRONS?
\
\
\
,
."
3.2 \\
3.0
Y
HADRONS
1 's0 11'+ 11'- ~411'
2.8
K+ K- K+ K-+211'
2.6 HADRONS? 2y
o 2,1,0
SPIN
;c
I
I
I
I
I
Table 3.1
SPEAR-LBL
DO 1863.6 2 MeV
D+ 1868.7 2 MeV
(D*)o 2006 1.5 HeV
8 5.1 2.8 } +
t1 + 8 HO
8* 2.6 2.5
DASP
F+ 2030 60 MeV
too far from the naive approach: counting the degrees of free-
dom. The virtual weak boson emitted by the decaying charmed
quark can either go into a lepton pair, W+ + (v e+) and
e
(~~ ~+), or into a quark pair, W+ + (du)red' (du)yellow'
(du)blue. Therefore we have 2 degrees of freedom for le?tonic
final states and 3 for hadronic ones (because of colour). This
implies that 40% of all decays should be semileptonic and 60%
hadronic. These naive quark-counting rules do not work for
the decay of strange particles. For example (A + S)/(A + all) ~
(D + evX) _ +
(D + all) - (11 - 3)%
(qqq)
\t,---~\,
(qq) (qq)
1182 A. ZICHICHI
2020! 3 MeV
30 ~
N
,---
2204!5 MeV
~
~
~ 20
0 t935!3 MeV
C\J
"- ~
if)
f-
Z
w 10
>
w
OL-~~~I~~__~____~~==__
2000
pp invariant mass
!l3998 events)
300
250
, 2.15t01 GeV
~
Q,
200
II>
C
-...
Q)
>
Q)
o 150
Q)
.J:l
E
::I
Z
100
50
O~ ____~______~______~____~________~
2.0 2.5 3.0 3.5 4.0
Table 3.2
( : g)
baryonium ,
I
C_....:...~_--.::~=----)
}
B o
mesonium
( ~ eee )
mesobaryon } B
1
pseudobaryon
c""----e_e_e___
e..:..e_e__) d 1Ob aryon 1st} B
2
( : ::) dibaryon
2nd
i) dissociation:
C: 0
0 ) .- ~
ii) simple cutting: ~
(: 0
0 ) C(V 8)
,
iii) rearrangements of (qq):
c: g)
(;) (;)
(: ~g) -----'"
STATUS OF THE SUBNUCLEAR WHYS 1185
iv) dissociation:
~
:)
~~
These mechanisms correspond to different decay rates: for
example, the "dissociation" is super-allowed. The exotic states
can decay into either a meson-meson or a baryon-meson system. Some
of them are expected to be narrow, others should be relatively
broad. If all these speculations are valid, a large set of new
states should show up.
u c
I
d
1s
(0 )
~\
ds b
For the sake of clarity, in Fig. 4.2 we show how the various mixing
angles come into the different quark transitions.
Notice that ~n this theory with six quarks, three mixing angles,
and a phase, the CP violating amplitude is in general depending on
the product
s. sin 8.
1 1
c. cos 8.
1 1
STATUS OF THE SUBNUCLEAR WHYS 1189
5. Do We Understand Leptons?
(5.1)
The best evidence for the new lepton T comes from (e~) final states
originally observed at SPEAR 14 ) and later at DORIS. The latest re-
sUlts 15 ) for the mass value are from two groups at DORIS:
T -)- IT\!
T
(9.00 2.9)%
T -)- all
Let me recall that the search for heavy leptons, exactly along the
line which has been found to be the cleanest, i.e. (~e) pairs in
the final state, has been developed originally in Frascati with
ADONE 16 ) . The machine energy was a few hundred HeV below threshold,
and the only result obtained on the (e~) pairs was a limit 17 ) on
the heavy lepton mass: Mh ~ 1 GeV.
Nw + SW
Qe m. = TW + ':":---..",.---'-_
3 2
Table 5.1
Values of intrinsic weak quantum numbers for left- and
right-handed leptons and quarks in the SU(2) x U(l)
Salam-Weinberg gauge theory of weak and electromagnetic
interactions
r:
(: t (;t
NW SW YN Qe.m.
1
- 1 +- 0 - 1 0
2
CIl
Z
(::t - 1 - -21 0 - 1 - 1
0
E-<
------------------------ ----- --- - ---- ----- ----
p.,
ril (e-)R (!n R (-r-)R - 1 0 - 1 - 2 - 1
....:I
(:1 (.J CJ
Ul 1 1 1 2
+- 0 +- +-
~ 3 2 3 3
1
1 1
- -31
0
u - 0 +-
3 -2 3
M
X --------------------------- r----- ----- ----- ----- -----
CIl
2
::.::
p:: (dC)R (sC)R (bC)R -
1
3
0 - 1 - 3 -"31
~
~ 1 4 2
G (u)R (c)R (t)R 0 + 1 - +-
"3 3 3
Notice that we have SlX leptons and SlX quarks, i.e. an equal
number of quarks and leptons. Why is this needed? It is needed
in order to avoid the Adler-Bell-Jackiw anomaly18) of the axial
vector currents, which spoils renormalizability. This anomaly
comes from diagrams of the type
where W3 lS the weak gauge boson coupled to the weak isospin (T~),
and B lS the weak gauge boson coupled to the weak hypercharge yW.
The arrowed lines indicate the basic fermions of the theory (lep-
tons and quarks). The basic Lagrangian (Salam-Weinberg)ll) is in
fact
!.t
e.m.+weak g[L: ~t ~ wt) . w+ g'(Li ~iywWi) B ,
i
where T is the weak isospin operator and yW the weak hyper charge
operator. The ratio of the two coupling constants
tan 8 g' / g
w
is the famous weak angle 8. Notice that this angle is nothing more
w
than the measurement of the mixing between the two symmetry groups
which put leptons and quarks on the same footing, SU(2) and U(l).
This mixing should obviously be the same for quarks and leptons.
In order to cancel the ABJ anomaly, the electric charges of all
leptons and quarks should be zero:
STATUS OF THE SUBNUCLEAR WHYS 1193
So, given six leptons we ought to have six quarks with the correct
electric charge states. If the electric charge of the quarks is
fractional, as in the standard Gell-Mann- Zweig model, three colours
are needed in order to get zero for the total charges of all lep-
tons plus quarks. In fact, with (e-, ~-, T-), the total lepton
charge is -3, while the total quark charge with (u d), (c s), (t b)
is (%-%) + (213-%) + (%-%) = +1, which becomes +3 if we intro-
duce three colours. The introduction of colour is essential if we
are to avoid the above-mentioned triangle anomalies.
This is not the end of the story. Suppose (we are physicists,
~.e. the facts are needed to establish the truth) that the process
coming into play is
+ V
e
+ V
~
1194 A. ZICHICHI
e + ~+ + neutrinos
e + e+ + neutrinos
~ + (hadrons)+ + neutrinos
e + (hadrons)+ + neutrinos
ee
~p
1.0 0.3 ,
ee
e~
~~
e~
0.5 0.3 ,
e + tracks
~ + tracks
1.0 0.3 .
Table 5.2
ee ee ww e + tracks
wii
-ew =
ew W + tracks
Experimental data 1 0.3 0.5 0.3 l.0 0.3
p
Heavy lepton with
0
.~
its own neutrino; 1 0.5 1
.j..J
ctt
as In SU(2) x U(l)
.j..J
u
Q)
p..
Two "gauge leptons"
~
Q)
E, H, quasi-degenerate 1 0.75 1
in masses
H
0
Q)
q
2/3 -+q%
.c + V + e 1 0.5 1
E-! e
2.0
1.6
A 1.2
I- 400
a
v 0.8
CPR
1
eAnderson et 01.
ee
CF -Hom et 01.
0.4
0
0 4 8 12
Q (GeV)
A e-bM
with
The datal;) show the same scale behaviour as (IN) inclusive produc-
tion (Fig. 3.1). A comparison with the data of Section 3 suggests
that (IN) and "continuum" corne from similar processes.
ii) For PT > 1 GeV/c the bulk of single-lepton production (> 75%)
is due to pairs. However, the production of known vector
mesons (p, w, ~) is not enough. An important contribution
from the continuum is needed.
1198 A. ZICHICHI
-32
10
~
.8CSY
N
CP-I
~
l?
-33 . .... CFS
N
I
E
10 .
0
u ~I't
.'+'.~
"
>-
>-
-34
.~
"U
~
10
M~
"U T
......
N b
II
"U
'"~
1035
-36
10
iii) The only firm evidence for prompt single leptons comes from a
recent Fermilab experiment where direct single-muon production
has been observed with PT < 1 GeV/c and at a rate comparable
to that of ~ pairs 19 ) . The origin of these prompt single muons
~s believed to be "charm" production.
Charged currents 20 )
Neutral currents 21 )
do
--
dy
= a + bel - y )
2
for v,
do 'iJ.,
-- = a(l for
dy
with a = 0.29 0.01; b = 0.06 0.01.
~
o
o
T T
1/
051- 8=0,8
--=~~=f=--
~.+ --+-
I I
>- 0.41-
.HPWF1
1.01- .. " " 2 -
CERN WA1
CERN -
0.31-
I I I
0,51- 1/
-
HFWF
WA 1
~ tit-~
~ 05~ k ~ I -+--+--+-+-+
f
~
~O+ !+~~=HI I~
0.3f- +-=F -
I I I 01 I I I 1
o 50 100 150 o 50 100 150 200
E-1/ Ev. GeV
Fig. 7.1 y-dependence versus neutrino or anti- Fig. 7.2 Comparison of the ratio between the
cross-sections of antineutrinos and neutrinos 1>
neutrino energy as measured at CERN (WAI Group) N
and Fermilab (HPWF) versus energy as measured at CERN (WAI Group) ()
and Fermilab (HPWF) :c
(')
:c
STATUS OF THE SUBNUCLEAR WHYS 1201
(n.>
I 3'
3 4 5 6 5 10 15 20
t:.y
Fig. 8.1 Asymmetry parameter as a function of Fig. 8.2 Showing the correlation between the
6y for a) unlike pairs, b) like pairs. Notice average charged multiplicities in the forward
that the asymmetry extends to very high values and backward hemispheres in (pp) collisions
of 6y; and this is the evidence for "long- at IS = 63 GeV
range" effects. o
'"
Co)
1204 A. ZICHICHI
and lose its chance to annihilate. Pairs with M!~ 1 GeV can be
made up with fast partons. These pass almost freely through the
nucleus and can annihilate anywhere in it; at the surface or deep
inside. The A-dependence of the lepton pair cross-section should
vary with A~3 for M ~ 1 GeV and become proportional to A for
M 1 GeV. The interpolation between these two limits depends on
the details of quark interactions. Notice that for large Pr ha-
. processes, such as the ab
dron1c '
ove-ment10ned ( n-K-p-
+ + +).1nc 1US1ve
.
Nothing new.
15
10
0:-
a
0
a.le
--
z 3
as
~
I /)
C
GI
E
.~
a.le
8.5
~
3
w
o .2 .4 .6 .8 1.0
(GeV/c)
Fig. 11.1 Showing the values of PT for all quark
experiments. Note that 16 measurements have been
made at PT = 0.1 GeV/c and none above PT = 1 GeV/c.
The latest results of Cronin et ale are not in-
cluded. See text.
1206 A. ZICHICHI
pute the transition from the virtual cloud state to the real one.
If this required, say, 20 GeV, the quarks could be very light;
still, in order to break a proton, a threshold energy of 20 GeV is
necessary.
No, if nature has at least six quarks and six leptons, as dis-
cussed in Section 4.
No.
Conclusions
th
It was the 15 year of the School, and we wanted to focus the
program on all the problems we could think of32): some hot and new;
others old and quasi-forgotten, in spite of their great relevance
to our understanding of the laws of nature.
STATUS OF THE SUBNUCLEAR WHYS 1211
References
12) P.W. Higgs, Phys. Lett. 12, 132 (1964); Phys. Rev. Lett. 11,
508 (1964); Phys. Rev. 145, 1156 (1966).
1215
1216 LIST OF PRIZES AND SCHOLARSHIPS
1217
1218 LIST OF PARTICIPANTS
1231
1232 INDEX